AHD Manuscript (test)
Joshua Fieggen
28/03/2025
Link to cohort
library(readxl)
library(tidyverse)
library(magrittr)
library(ggpubr)
library(cowplot)
library(ggsci)
library(lubridate)
library(zoo)
library(formattable)
library(caret)
library(plotROC)
library(dplyr)
library(ggplot2)
library(shiny)
library(readr)
library(eeptools)
library(survival)
library(survminer)
library(EpiStats)
library(tableone)
library(gtsummary)
setwd("/Volumes/Extreme SSD/Survival final")
Cohort <- read_csv("Cohort_Final_test.csv")Summary of the cohort and exploring data
## study_id Enumeration_date Censored_Date Censored
## Min. :1.001e+09 Min. :2017-01-03 Min. :2017-01-12 Min. :0.0000
## 1st Qu.:3.194e+09 1st Qu.:2017-08-30 1st Qu.:2021-03-31 1st Qu.:0.0000
## Median :5.430e+09 Median :2018-07-30 Median :2021-03-31 Median :0.0000
## Mean :5.441e+09 Mean :2018-09-21 Mean :2021-01-24 Mean :0.1011
## 3rd Qu.:7.664e+09 3rd Qu.:2019-08-22 3rd Qu.:2021-03-31 3rd Qu.:0.0000
## Max. :9.998e+09 Max. :2021-03-30 Max. :2021-08-05 Max. :1.0000
##
## Linked Time Enumeration_CD4 Enumeration_Facility
## Min. :0.0000 Min. : 0.0 Min. : 0.0 Length:13344
## 1st Qu.:1.0000 1st Qu.: 489.8 1st Qu.: 57.0 Class :character
## Median :1.0000 Median : 895.5 Median :111.0 Mode :character
## Mean :0.8357 Mean : 856.1 Mean :106.9
## 3rd Qu.:1.0000 3rd Qu.:1272.0 3rd Qu.:158.0
## Max. :1.0000 Max. :1620.0 Max. :199.0
##
## Latest_CD4 Enum_CD4_Cat Enumeration_CD4_Cat E_CD4_Cat
## Min. : 0.0 Length:13344 Length:13344 Min. :0.0000
## 1st Qu.: 60.0 Class :character Class :character 1st Qu.:0.0000
## Median :116.0 Mode :character Mode :character Median :0.0000
## Mean :109.5 Mean :0.6654
## 3rd Qu.:162.0 3rd Qu.:1.0000
## Max. :199.0 Max. :2.0000
##
## Hospital_Enumerations VL_Enum VL_Latest VL_E_cat
## Min. :0.0000 Min. : 0 Min. : 0 Length:13344
## 1st Qu.:0.0000 1st Qu.: 0 1st Qu.: 0 Class :character
## Median :0.0000 Median : 4420 Median : 0 Mode :character
## Mean :0.1546 Mean : 157778 Mean : 53363
## 3rd Qu.:0.0000 3rd Qu.: 79222 3rd Qu.: 858
## Max. :1.0000 Max. :48267460 Max. :8484391
## NA's :9310 NA's :3869
## VL_E_Cat VL_E_Not_Done VL_L_cat VL_L_Cat
## Min. :0.000 Min. :0.0000 Length:13344 Min. :0.000
## 1st Qu.:0.000 1st Qu.:0.0000 Class :character 1st Qu.:0.000
## Median :2.000 Median :1.0000 Mode :character Median :1.000
## Mean :1.243 Mean :0.6977 Mean :1.294
## 3rd Qu.:2.000 3rd Qu.:1.0000 3rd Qu.:3.000
## Max. :2.000 Max. :1.0000 Max. :3.000
## NA's :9310
## Incident_Crypto Blood_Pos_I Reflex_Pos_I Prevalent_Crypto
## Min. :0.00000 Min. :0.00 Min. :0.000 Min. :0.00000
## 1st Qu.:0.00000 1st Qu.:1.00 1st Qu.:1.000 1st Qu.:0.00000
## Median :0.00000 Median :1.00 Median :1.000 Median :0.00000
## Mean :0.01589 Mean :0.92 Mean :0.783 Mean :0.01686
## 3rd Qu.:0.00000 3rd Qu.:1.00 3rd Qu.:1.000 3rd Qu.:0.00000
## Max. :1.00000 Max. :1.00 Max. :1.000 Max. :1.00000
## NA's :13132 NA's :13132
## Blood_Pos_P Reflex_Pos_P Previous_Crypto Blood_Pos_Prev
## Min. :0.000 Min. :0.000 Min. :0.000000 Min. :0.000
## 1st Qu.:1.000 1st Qu.:1.000 1st Qu.:0.000000 1st Qu.:0.000
## Median :1.000 Median :1.000 Median :0.000000 Median :1.000
## Mean :0.947 Mean :0.889 Mean :0.008918 Mean :0.672
## 3rd Qu.:1.000 3rd Qu.:1.000 3rd Qu.:0.000000 3rd Qu.:1.000
## Max. :1.000 Max. :1.000 Max. :1.000000 Max. :1.000
## NA's :13119 NA's :13119 NA's :13225
## Reflex_Pos_Prev Reflex_LFA_Done Reflex_LFA_Positive first_ART_date
## Min. :0.000 Min. :0.0000 Min. :1 Min. :2000-01-03
## 1st Qu.:0.000 1st Qu.:0.0000 1st Qu.:1 1st Qu.:2012-10-16
## Median :0.000 Median :1.0000 Median :1 Median :2016-06-15
## Mean :0.244 Mean :0.5127 Mean :1 Mean :2015-06-18
## 3rd Qu.:0.000 3rd Qu.:1.0000 3rd Qu.:1 3rd Qu.:2018-08-10
## Max. :1.000 Max. :1.0000 Max. :1 Max. :2021-03-30
## NA's :13225 NA's :13015 NA's :249
## recent_ART ART_Exp_C Subs_Initiated Date_First_Initiated
## Mode :logical Length:13344 Min. :0.000 Min. :2017-01-03
## FALSE:10713 Class :character 1st Qu.:1.000 1st Qu.:2017-11-14
## TRUE :2631 Mode :character Median :1.000 Median :2018-11-13
## Mean :0.953 Mean :2018-11-26
## 3rd Qu.:1.000 3rd Qu.:2019-10-09
## Max. :1.000 Max. :2021-03-30
## NA's :7659 NA's :7925
## Time_To_First_ART Reinitiated Date_Reinitiated Time_Reinitiation
## Min. : 0.00 Min. :0.000 Min. :2017-01-03 Min. : 0.0
## 1st Qu.: 0.00 1st Qu.:1.000 1st Qu.:2017-11-09 1st Qu.: 0.0
## Median : 10.00 Median :1.000 Median :2018-10-08 Median : 7.0
## Mean : 49.62 Mean :0.919 Mean :2018-11-14 Mean : 39.4
## 3rd Qu.: 28.00 3rd Qu.:1.000 3rd Qu.:2019-10-04 3rd Qu.: 25.0
## Max. :1503.00 Max. :1.000 Max. :2021-03-31 Max. :1287.0
## NA's :7925 NA's :8064 NA's :8491 NA's :8491
## DOB Sex Age AgeCat
## Min. :1929-07-06 Min. :1.000 Min. :18.01 Length:13344
## 1st Qu.:1975-12-27 1st Qu.:1.000 1st Qu.:30.34 Class :character
## Median :1982-11-17 Median :2.000 Median :35.78 Mode :character
## Mean :1981-08-22 Mean :1.571 Mean :37.08
## 3rd Qu.:1988-06-04 3rd Qu.:2.000 3rd Qu.:42.70
## Max. :2002-09-28 Max. :2.000 Max. :89.25
##
## Age_Cat Age_cat Sex_B NPR_Vital_Status
## Min. :0.0000 Length:13344 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.0000 Class :character 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.0000 Mode :character Median :0.0000 Median :0.0000
## Mean :0.3375 Mean :0.4293 Mean :0.1021
## 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:0.0000
## Max. :1.0000 Max. :1.0000 Max. :1.0000
##
## NPR_Date Current_Diabetes Current_Hypertension
## Min. :2017-01-12 00:00:00.00 Min. :0.00000 Min. :0.00000
## 1st Qu.:2018-06-23 12:00:00.00 1st Qu.:0.00000 1st Qu.:0.00000
## Median :2019-07-10 00:00:00.00 Median :0.00000 Median :0.00000
## Mean :2019-06-22 15:16:39.11 Mean :0.02608 Mean :0.08933
## 3rd Qu.:2020-06-29 00:00:00.00 3rd Qu.:0.00000 3rd Qu.:0.00000
## Max. :2021-08-05 00:00:00.00 Max. :1.00000 Max. :1.00000
## NA's :11982
## Current_CKD Current_AKI Previous_AKI Incident_AKI
## Min. :0.00000 Min. :0.0000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.00000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.00000 Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.01963 Mean :0.1212 Mean :0.1318 Mean :0.1418
## 3rd Qu.:0.00000 3rd Qu.:0.0000 3rd Qu.:0.0000 3rd Qu.:0.0000
## Max. :1.00000 Max. :1.0000 Max. :1.0000 Max. :1.0000
##
## Current_TB_PHDC Previous_TB_PHDC Incident_TB_PHDC
## Min. :0.0000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.0000 Median :0.0000 Median :0.0000
## Mean :0.2548 Mean :0.3293 Mean :0.1658
## 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:0.0000
## Max. :1.0000 Max. :1.0000 Max. :1.0000
## make changes to data structure and lables
#Changes to Data
Cohort$Enumeration_CD4 <- as.numeric(Cohort$Enumeration_CD4)
Data <- select(Cohort, study_id, Censored, Time, Enumeration_CD4, VL_E_Cat,VL_E_Not_Done, E_CD4_Cat, Hospital_Enumerations,
Incident_Crypto, Prevalent_Crypto, Previous_Crypto,
ART_Exp_C, Subs_Initiated, Time_To_First_ART, Reinitiated, Time_Reinitiation
, Sex_B, Age, Age_Cat, Linked, Current_Diabetes, Current_Hypertension,
Current_CKD, Current_TB_PHDC, Previous_TB_PHDC, Incident_TB_PHDC)
Data <- within(Data, {
`Viral Load at Enumeration` <- factor(VL_E_Cat, labels = c("< 100", "100 -1000", "> 1000"))
`Enumeration VL Not Done` <- VL_E_Not_Done
ART <- factor(ART_Exp_C, levels = c("On ART","LTFU","Naive"), labels = c("On ART","LTFU","ART Naive"))
LTFU <- factor(Reinitiated, labels = c("Never Re-initiated", "Re-initiated"))
`ART Naive` <- factor(Subs_Initiated, labels = c("Never Initiaited", "Subsequently Initiated"))
`Enumeration CD4` <- Enumeration_CD4
`Enumeration CD4 Categorical` <- factor(E_CD4_Cat, labels = c("101 - 199", "51 - 100", "0 - 50"))
`Time in Cohort` <- Time
`Time to Initiation` <- Time_To_First_ART
`Time to Re-initiation` <- Time_Reinitiation
Mortality <- factor(Censored, labels = c("Survived", "Died"))
Sex <- factor(Sex_B, labels = c("Female", "Male"))
`Sex (Male)` <- Sex_B
`Age > 40` <- Age_Cat
`Age` <- Age
`Incident Cryptococcosis` <- Incident_Crypto
`Current Cryptococcosis` <- Prevalent_Crypto
`Previous Cryptococcosis` <- Previous_Crypto
`Enumerated in Hospital` <- Hospital_Enumerations
`NPR Linkage` <- factor(Linked, labels = c("Unlinked", "Linked"))
`Diabetic` <- Current_Diabetes
`Hypertension` <- Current_Hypertension
`Chronic Kidney Disease` <- Current_CKD
`Current TB (PHDC)` <- Current_TB_PHDC
`Previous TB (PHDC)` <- Previous_TB_PHDC
`Incident TB (PHDC)` <- Incident_TB_PHDC})
Data <- select(Data, study_id, `Time`, `Viral Load at Enumeration`, `Enumeration VL Not Done`,
`Enumeration CD4`, `Enumerated in Hospital`, `Enumeration CD4 Categorical`,
`NPR Linkage`, `Mortality`,
`Sex`, `Sex (Male)`, `Age > 40`, `Age`,
`Incident TB (PHDC)`, `Previous TB (PHDC)`, `Current TB (PHDC)`,
`Previous Cryptococcosis`, `Current Cryptococcosis`, `Incident Cryptococcosis`,
`Chronic Kidney Disease`, `Hypertension`, `Diabetic`,
`ART Naive`, `LTFU`, `ART`,
`Time to Initiation`, `Time to Re-initiation`, Censored, VL_E_Cat)
colnames(Data)## [1] "study_id" "Time"
## [3] "Viral Load at Enumeration" "Enumeration VL Not Done"
## [5] "Enumeration CD4" "Enumerated in Hospital"
## [7] "Enumeration CD4 Categorical" "NPR Linkage"
## [9] "Mortality" "Sex"
## [11] "Sex (Male)" "Age > 40"
## [13] "Age" "Incident TB (PHDC)"
## [15] "Previous TB (PHDC)" "Current TB (PHDC)"
## [17] "Previous Cryptococcosis" "Current Cryptococcosis"
## [19] "Incident Cryptococcosis" "Chronic Kidney Disease"
## [21] "Hypertension" "Diabetic"
## [23] "ART Naive" "LTFU"
## [25] "ART" "Time to Initiation"
## [27] "Time to Re-initiation" "Censored"
## [29] "VL_E_Cat"Data$AgeCat_4<-cut(Data$Age, c(0,25,35,45,100))
summary(Data$AgeCat_4)## (0,25] (25,35] (35,45] (45,100]
## 1031 5154 4637 2522Data$Age_Cat_4 <- NA
Data$Age_Cat_4[Data$AgeCat_4=="(0,25]"] <- 0
Data$Age_Cat_4[Data$AgeCat_4=="(25,35]"] <- 1
Data$Age_Cat_4[Data$AgeCat_4=="(35,45]"] <- 2
Data$Age_Cat_4[Data$AgeCat_4=="(45,100]"] <- 3
table(Data$Age_Cat_4)##
## 0 1 2 3
## 1031 5154 4637 2522names(Data)[30] <- "Age_cat"
Data <- within(Data, {Age_cat <- factor(Age_cat, labels = c("18-25", "25-35", "35-45", ">45"))})
Data$Age_cat = relevel(Data$Age_cat, ref = "25-35")
summary(Data$Age_cat)## 25-35 18-25 35-45 >45
## 5154 1031 4637 2522Plot by mortality
Mort_Plots <- subset(Cohort, Linked > 0)
Mort_Plots <- within(Mort_Plots, { Mortality <- factor(Censored, labels = c("Survived", "Died"))})
ggdensity(Mort_Plots, x = "Enumeration_CD4",
add = "median",
color = "Mortality", fill = "Mortality",
palette = c("#00AFBB", "#E7B800"))
ggdensity(Mort_Plots, x = "Time",
add = "median",
color = "Mortality", fill = "Mortality",
palette = c("#00AFBB", "#E7B800"))
ggdensity(Mort_Plots, x = "Time_To_First_ART",
add = "median",
color = "Mortality", fill = "Mortality",
palette = c("#00AFBB", "#E7B800"))## Warning: Removed 6627 rows containing non-finite outside the scale range
## (`stat_density()`).
ggdensity(Mort_Plots, x = "Time_Reinitiation",
add = "median",
color = "Mortality", fill = "Mortality",
palette = c("#00AFBB", "#E7B800"))## Warning: Removed 7035 rows containing non-finite outside the scale range
## (`stat_density()`).
ggdensity(Mort_Plots, x = "Age",
add = "median",
color = "Mortality", fill = "Mortality",
palette = c("#00AFBB", "#E7B800"))
Plot by ART
Create stratified summary tables
## Table_1
library(tableone)
CreateTableOne(data = Data)##
## Overall
## n 13344
## study_id (mean (SD)) 5441175670.67 (2587845360.28)
## Time (mean (SD)) 856.07 (463.40)
## Viral Load at Enumeration (%)
## < 100 1311 (32.5)
## 100 -1000 431 (10.7)
## > 1000 2292 (56.8)
## Enumeration VL Not Done (mean (SD)) 0.70 (0.46)
## Enumeration CD4 (mean (SD)) 106.89 (57.78)
## Enumerated in Hospital (mean (SD)) 0.15 (0.36)
## Enumeration CD4 Categorical (%)
## 101 - 199 7368 (55.2)
## 51 - 100 3073 (23.0)
## 0 - 50 2903 (21.8)
## NPR Linkage = Linked (%) 11152 (83.6)
## Mortality = Died (%) 1349 (10.1)
## Sex = Male (%) 5728 (42.9)
## Sex (Male) (mean (SD)) 0.43 (0.49)
## Age > 40 (mean (SD)) 0.34 (0.47)
## Age (mean (SD)) 37.08 (9.45)
## Incident TB (PHDC) (mean (SD)) 0.17 (0.37)
## Previous TB (PHDC) (mean (SD)) 0.33 (0.47)
## Current TB (PHDC) (mean (SD)) 0.25 (0.44)
## Previous Cryptococcosis (mean (SD)) 0.01 (0.09)
## Current Cryptococcosis (mean (SD)) 0.02 (0.13)
## Incident Cryptococcosis (mean (SD)) 0.02 (0.13)
## Chronic Kidney Disease (mean (SD)) 0.02 (0.14)
## Hypertension (mean (SD)) 0.09 (0.29)
## Diabetic (mean (SD)) 0.03 (0.16)
## ART Naive = Subsequently Initiated (%) 5419 (95.3)
## LTFU = Re-initiated (%) 4853 (91.9)
## ART (%)
## On ART 2379 (17.8)
## LTFU 5280 (39.6)
## ART Naive 5685 (42.6)
## Time to Initiation (mean (SD)) 49.62 (140.82)
## Time to Re-initiation (mean (SD)) 39.41 (110.89)
## Censored (mean (SD)) 0.10 (0.30)
## VL_E_Cat (mean (SD)) 1.24 (0.91)
## Age_cat (%)
## 25-35 5154 (38.6)
## 18-25 1031 ( 7.7)
## 35-45 4637 (34.7)
## >45 2522 (18.9)
## Age_Cat_4 (mean (SD)) 1.65 (0.87)myVars <- c("Mortality", "NPR Linkage", "Sex", "Age > 40", "Age", "Enumeration CD4", "Enumeration CD4 Categorical",
"Time in Cohort",
"Viral Load at Enumeration", "Enumeration VL Not Done",
"ART", "ART Naive", "Time to Initiation", "LTFU", "Time to Re-initiation",
"Enumerated in Hospital",
"Current TB (PHDC)", "Previous TB (PHDC)", "Incident TB (PHDC)",
"Current Cryptococcosis", "Previous Cryptococcosis", "Incident Cryptococcosis","Age_cat")
catVars <- c("Mortality", "NPR Linkage", "Sex", "Age > 40", "Enumeration CD4 Categorical",
"Viral Load at Enumeration", "Enumeration VL Not Done",
"Enumerated in Hospital", "Current TB (PHDC)", "Previous TB (PHDC)", "Incident TB (PHDC)",
"ART", "ART Naive", "LTFU",
"Previous Cryptococcosis", "Current Cryptococcosis", "Incident Cryptococcosis","Age_cat")
tab2 <- CreateTableOne(vars = myVars, data = Data, factorVars = catVars)
#Add non-normality
non_normal <- c("Age", "Time to Re-initiation", "Time to Initiation", "Time in Cohort", "Enumeration CD4")
print(tab2, nonnormal = non_normal, quote = TRUE, noSpaces = TRUE)## ""
## "" "Overall"
## "n" "13344"
## "Mortality = Died (%)" "1349 (10.1)"
## "NPR Linkage = Linked (%)" "11152 (83.6)"
## "Sex = Male (%)" "5728 (42.9)"
## "Age > 40 = 1 (%)" "4503 (33.7)"
## "Age (median [IQR])" "35.78 [30.34, 42.70]"
## "Enumeration CD4 (median [IQR])" "111.00 [57.00, 158.00]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "7368 (55.2)"
## " 51 - 100" "3073 (23.0)"
## " 0 - 50" "2903 (21.8)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "1311 (32.5)"
## " 100 -1000" "431 (10.7)"
## " > 1000" "2292 (56.8)"
## "Enumeration VL Not Done = 1 (%)" "9310 (69.8)"
## "ART (%)" ""
## " On ART" "2379 (17.8)"
## " LTFU" "5280 (39.6)"
## " ART Naive" "5685 (42.6)"
## "ART Naive = Subsequently Initiated (%)" "5419 (95.3)"
## "Time to Initiation (median [IQR])" "10.00 [0.00, 28.00]"
## "LTFU = Re-initiated (%)" "4853 (91.9)"
## "Time to Re-initiation (median [IQR])" "7.00 [0.00, 25.00]"
## "Enumerated in Hospital = 1 (%)" "2063 (15.5)"
## "Current TB (PHDC) = 1 (%)" "3400 (25.5)"
## "Previous TB (PHDC) = 1 (%)" "4394 (32.9)"
## "Incident TB (PHDC) = 1 (%)" "2212 (16.6)"
## "Current Cryptococcosis = 1 (%)" "225 (1.7)"
## "Previous Cryptococcosis = 1 (%)" "119 (0.9)"
## "Incident Cryptococcosis = 1 (%)" "212 (1.6)"
## "Age_cat (%)" ""
## " 25-35" "5154 (38.6)"
## " 18-25" "1031 (7.7)"
## " 35-45" "4637 (34.7)"
## " >45" "2522 (18.9)"#Linked vs unlinked
tab_linked <- CreateTableOne(vars = myVars, strata = "NPR Linkage" , data = Data, factorVars = catVars)
print(tab_linked, nonnormal = non_normal, quote = TRUE, noSpaces = TRUE)## "Stratified by NPR Linkage"
## "" "Unlinked"
## "n" "2192"
## "Mortality = Died (%)" "0 (0.0)"
## "NPR Linkage = Linked (%)" "0 (0.0)"
## "Sex = Male (%)" "1024 (46.7)"
## "Age > 40 = 1 (%)" "863 (39.4)"
## "Age (median [IQR])" "37.24 [31.51, 44.35]"
## "Enumeration CD4 (median [IQR])" "107.00 [56.00, 157.25]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "1176 (53.6)"
## " 51 - 100" "520 (23.7)"
## " 0 - 50" "496 (22.6)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "211 (29.6)"
## " 100 -1000" "66 (9.2)"
## " > 1000" "437 (61.2)"
## "Enumeration VL Not Done = 1 (%)" "1478 (67.4)"
## "ART (%)" ""
## " On ART" "422 (19.3)"
## " LTFU" "813 (37.1)"
## " ART Naive" "957 (43.7)"
## "ART Naive = Subsequently Initiated (%)" "894 (93.4)"
## "Time to Initiation (median [IQR])" "12.00 [0.25, 27.00]"
## "LTFU = Re-initiated (%)" "736 (90.5)"
## "Time to Re-initiation (median [IQR])" "5.00 [0.00, 21.00]"
## "Enumerated in Hospital = 1 (%)" "393 (17.9)"
## "Current TB (PHDC) = 1 (%)" "611 (27.9)"
## "Previous TB (PHDC) = 1 (%)" "831 (37.9)"
## "Incident TB (PHDC) = 1 (%)" "351 (16.0)"
## "Current Cryptococcosis = 1 (%)" "40 (1.8)"
## "Previous Cryptococcosis = 1 (%)" "27 (1.2)"
## "Incident Cryptococcosis = 1 (%)" "36 (1.6)"
## "Age_cat (%)" ""
## " 25-35" "753 (34.4)"
## " 18-25" "136 (6.2)"
## " 35-45" "806 (36.8)"
## " >45" "497 (22.7)"
## "Stratified by NPR Linkage"
## "" "Linked" "p"
## "n" "11152" ""
## "Mortality = Died (%)" "1349 (12.1)" "<0.001"
## "NPR Linkage = Linked (%)" "11152 (100.0)" "<0.001"
## "Sex = Male (%)" "4704 (42.2)" "<0.001"
## "Age > 40 = 1 (%)" "3640 (32.6)" "<0.001"
## "Age (median [IQR])" "35.54 [30.16, 42.36]" "<0.001"
## "Enumeration CD4 (median [IQR])" "111.00 [58.00, 158.00]" "0.400"
## "Enumeration CD4 Categorical (%)" "" "0.268"
## " 101 - 199" "6192 (55.5)" ""
## " 51 - 100" "2553 (22.9)" ""
## " 0 - 50" "2407 (21.6)" ""
## "Viral Load at Enumeration (%)" "" "0.031"
## " < 100" "1100 (33.1)" ""
## " 100 -1000" "365 (11.0)" ""
## " > 1000" "1855 (55.9)" ""
## "Enumeration VL Not Done = 1 (%)" "7832 (70.2)" "0.010"
## "ART (%)" "" "0.021"
## " On ART" "1957 (17.5)" ""
## " LTFU" "4467 (40.1)" ""
## " ART Naive" "4728 (42.4)" ""
## "ART Naive = Subsequently Initiated (%)" "4525 (95.7)" "0.003"
## "Time to Initiation (median [IQR])" "10.00 [0.00, 28.00]" "0.398"
## "LTFU = Re-initiated (%)" "4117 (92.2)" "0.133"
## "Time to Re-initiation (median [IQR])" "7.00 [0.00, 26.00]" "0.033"
## "Enumerated in Hospital = 1 (%)" "1670 (15.0)" "0.001"
## "Current TB (PHDC) = 1 (%)" "2789 (25.0)" "0.005"
## "Previous TB (PHDC) = 1 (%)" "3563 (31.9)" "<0.001"
## "Incident TB (PHDC) = 1 (%)" "1861 (16.7)" "0.456"
## "Current Cryptococcosis = 1 (%)" "185 (1.7)" "0.645"
## "Previous Cryptococcosis = 1 (%)" "92 (0.8)" "0.084"
## "Incident Cryptococcosis = 1 (%)" "176 (1.6)" "0.900"
## "Age_cat (%)" "" "<0.001"
## " 25-35" "4401 (39.5)" ""
## " 18-25" "895 (8.0)" ""
## " 35-45" "3831 (34.4)" ""
## " >45" "2025 (18.2)" ""
## "Stratified by NPR Linkage"
## "" "test"
## "n" ""
## "Mortality = Died (%)" ""
## "NPR Linkage = Linked (%)" ""
## "Sex = Male (%)" ""
## "Age > 40 = 1 (%)" ""
## "Age (median [IQR])" "nonnorm"
## "Enumeration CD4 (median [IQR])" "nonnorm"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" ""
## " 51 - 100" ""
## " 0 - 50" ""
## "Viral Load at Enumeration (%)" ""
## " < 100" ""
## " 100 -1000" ""
## " > 1000" ""
## "Enumeration VL Not Done = 1 (%)" ""
## "ART (%)" ""
## " On ART" ""
## " LTFU" ""
## " ART Naive" ""
## "ART Naive = Subsequently Initiated (%)" ""
## "Time to Initiation (median [IQR])" "nonnorm"
## "LTFU = Re-initiated (%)" ""
## "Time to Re-initiation (median [IQR])" "nonnorm"
## "Enumerated in Hospital = 1 (%)" ""
## "Current TB (PHDC) = 1 (%)" ""
## "Previous TB (PHDC) = 1 (%)" ""
## "Incident TB (PHDC) = 1 (%)" ""
## "Current Cryptococcosis = 1 (%)" ""
## "Previous Cryptococcosis = 1 (%)" ""
## "Incident Cryptococcosis = 1 (%)" ""
## "Age_cat (%)" ""
## " 25-35" ""
## " 18-25" ""
## " 35-45" ""
## " >45" ""#Stratified (Mortality status)
Linked_Data <- subset(Data, `NPR Linkage` == "Linked")
tab3 <- CreateTableOne(vars = myVars, strata = "Mortality" , data = Linked_Data, factorVars = catVars)
print(tab3, nonnormal = non_normal, quote = TRUE, noSpaces = TRUE)## "Stratified by Mortality"
## "" "Survived"
## "n" "9803"
## "Mortality = Died (%)" "0 (0.0)"
## "NPR Linkage = Linked (%)" "9803 (100.0)"
## "Sex = Male (%)" "4069 (41.5)"
## "Age > 40 = 1 (%)" "3080 (31.4)"
## "Age (median [IQR])" "35.27 [29.96, 41.86]"
## "Enumeration CD4 (median [IQR])" "117.00 [64.00, 161.00]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "5692 (58.1)"
## " 51 - 100" "2225 (22.7)"
## " 0 - 50" "1886 (19.2)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "947 (34.3)"
## " 100 -1000" "318 (11.5)"
## " > 1000" "1493 (54.1)"
## "Enumeration VL Not Done = 1 (%)" "7045 (71.9)"
## "ART (%)" ""
## " On ART" "1611 (16.4)"
## " LTFU" "3859 (39.4)"
## " ART Naive" "4333 (44.2)"
## "ART Naive = Subsequently Initiated (%)" "4181 (96.5)"
## "Time to Initiation (median [IQR])" "9.00 [0.00, 28.00]"
## "LTFU = Re-initiated (%)" "3617 (93.7)"
## "Time to Re-initiation (median [IQR])" "6.00 [0.00, 25.00]"
## "Enumerated in Hospital = 1 (%)" "1248 (12.7)"
## "Current TB (PHDC) = 1 (%)" "2295 (23.4)"
## "Previous TB (PHDC) = 1 (%)" "2902 (29.6)"
## "Incident TB (PHDC) = 1 (%)" "1533 (15.6)"
## "Current Cryptococcosis = 1 (%)" "135 (1.4)"
## "Previous Cryptococcosis = 1 (%)" "75 (0.8)"
## "Incident Cryptococcosis = 1 (%)" "126 (1.3)"
## "Age_cat (%)" ""
## " 25-35" "3976 (40.6)"
## " 18-25" "805 (8.2)"
## " 35-45" "3350 (34.2)"
## " >45" "1672 (17.1)"
## "Stratified by Mortality"
## "" "Died" "p"
## "n" "1349" ""
## "Mortality = Died (%)" "1349 (100.0)" "<0.001"
## "NPR Linkage = Linked (%)" "1349 (100.0)" "NA"
## "Sex = Male (%)" "635 (47.1)" "<0.001"
## "Age > 40 = 1 (%)" "560 (41.5)" "<0.001"
## "Age (median [IQR])" "37.70 [32.03, 45.39]" "<0.001"
## "Enumeration CD4 (median [IQR])" "69.00 [30.00, 128.00]" "<0.001"
## "Enumeration CD4 Categorical (%)" "" "<0.001"
## " 101 - 199" "500 (37.1)" ""
## " 51 - 100" "328 (24.3)" ""
## " 0 - 50" "521 (38.6)" ""
## "Viral Load at Enumeration (%)" "" "<0.001"
## " < 100" "153 (27.2)" ""
## " 100 -1000" "47 (8.4)" ""
## " > 1000" "362 (64.4)" ""
## "Enumeration VL Not Done = 1 (%)" "787 (58.3)" "<0.001"
## "ART (%)" "" "<0.001"
## " On ART" "346 (25.6)" ""
## " LTFU" "608 (45.1)" ""
## " ART Naive" "395 (29.3)" ""
## "ART Naive = Subsequently Initiated (%)" "344 (87.1)" "<0.001"
## "Time to Initiation (median [IQR])" "15.00 [4.00, 31.00]" "<0.001"
## "LTFU = Re-initiated (%)" "500 (82.2)" "<0.001"
## "Time to Re-initiation (median [IQR])" "8.00 [1.00, 27.00]" "0.006"
## "Enumerated in Hospital = 1 (%)" "422 (31.3)" "<0.001"
## "Current TB (PHDC) = 1 (%)" "494 (36.6)" "<0.001"
## "Previous TB (PHDC) = 1 (%)" "661 (49.0)" "<0.001"
## "Incident TB (PHDC) = 1 (%)" "328 (24.3)" "<0.001"
## "Current Cryptococcosis = 1 (%)" "50 (3.7)" "<0.001"
## "Previous Cryptococcosis = 1 (%)" "17 (1.3)" "0.085"
## "Incident Cryptococcosis = 1 (%)" "50 (3.7)" "<0.001"
## "Age_cat (%)" "" "<0.001"
## " 25-35" "425 (31.5)" ""
## " 18-25" "90 (6.7)" ""
## " 35-45" "481 (35.7)" ""
## " >45" "353 (26.2)" ""
## "Stratified by Mortality"
## "" "test"
## "n" ""
## "Mortality = Died (%)" ""
## "NPR Linkage = Linked (%)" ""
## "Sex = Male (%)" ""
## "Age > 40 = 1 (%)" ""
## "Age (median [IQR])" "nonnorm"
## "Enumeration CD4 (median [IQR])" "nonnorm"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" ""
## " 51 - 100" ""
## " 0 - 50" ""
## "Viral Load at Enumeration (%)" ""
## " < 100" ""
## " 100 -1000" ""
## " > 1000" ""
## "Enumeration VL Not Done = 1 (%)" ""
## "ART (%)" ""
## " On ART" ""
## " LTFU" ""
## " ART Naive" ""
## "ART Naive = Subsequently Initiated (%)" ""
## "Time to Initiation (median [IQR])" "nonnorm"
## "LTFU = Re-initiated (%)" ""
## "Time to Re-initiation (median [IQR])" "nonnorm"
## "Enumerated in Hospital = 1 (%)" ""
## "Current TB (PHDC) = 1 (%)" ""
## "Previous TB (PHDC) = 1 (%)" ""
## "Incident TB (PHDC) = 1 (%)" ""
## "Current Cryptococcosis = 1 (%)" ""
## "Previous Cryptococcosis = 1 (%)" ""
## "Incident Cryptococcosis = 1 (%)" ""
## "Age_cat (%)" ""
## " 25-35" ""
## " 18-25" ""
## " 35-45" ""
## " >45" ""#Stratified table 2 (LTFU/Re-initiation status)
tab4 <- CreateTableOne(vars = myVars, strata = "LTFU" , data = Data, factorVars = catVars)
print(tab4, nonnormal = non_normal, quote = TRUE, noSpaces = TRUE)## "Stratified by LTFU"
## "" "Never Re-initiated"
## "n" "427"
## "Mortality = Died (%)" "108 (25.3)"
## "NPR Linkage = Linked (%)" "350 (82.0)"
## "Sex = Male (%)" "168 (39.3)"
## "Age > 40 = 1 (%)" "161 (37.7)"
## "Age (median [IQR])" "37.33 [31.12, 43.60]"
## "Enumeration CD4 (median [IQR])" "95.00 [42.50, 157.00]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "201 (47.1)"
## " 51 - 100" "104 (24.4)"
## " 0 - 50" "122 (28.6)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "24 (17.1)"
## " 100 -1000" "19 (13.6)"
## " > 1000" "97 (69.3)"
## "Enumeration VL Not Done = 1 (%)" "287 (67.2)"
## "ART (%)" ""
## " On ART" "0 (0.0)"
## " LTFU" "427 (100.0)"
## " ART Naive" "0 (0.0)"
## "ART Naive = Subsequently Initiated (%)" "0 (NaN)"
## "Time to Initiation (median [IQR])" "NA [NA, NA]"
## "LTFU = Re-initiated (%)" "0 (0.0)"
## "Time to Re-initiation (median [IQR])" "NA [NA, NA]"
## "Enumerated in Hospital = 1 (%)" "194 (45.4)"
## "Current TB (PHDC) = 1 (%)" "121 (28.3)"
## "Previous TB (PHDC) = 1 (%)" "213 (49.9)"
## "Incident TB (PHDC) = 1 (%)" "35 (8.2)"
## "Current Cryptococcosis = 1 (%)" "15 (3.5)"
## "Previous Cryptococcosis = 1 (%)" "6 (1.4)"
## "Incident Cryptococcosis = 1 (%)" "1 (0.2)"
## "Age_cat (%)" ""
## " 25-35" "140 (32.8)"
## " 18-25" "31 (7.3)"
## " 35-45" "170 (39.8)"
## " >45" "86 (20.1)"
## "Stratified by LTFU"
## "" "Re-initiated" "p"
## "n" "4853" ""
## "Mortality = Died (%)" "500 (10.3)" "<0.001"
## "NPR Linkage = Linked (%)" "4117 (84.8)" "0.133"
## "Sex = Male (%)" "1820 (37.5)" "0.483"
## "Age > 40 = 1 (%)" "1592 (32.8)" "0.045"
## "Age (median [IQR])" "35.90 [30.69, 42.37]" "0.049"
## "Enumeration CD4 (median [IQR])" "107.00 [53.00, 155.00]" "0.086"
## "Enumeration CD4 Categorical (%)" "" "0.026"
## " 101 - 199" "2583 (53.2)" ""
## " 51 - 100" "1131 (23.3)" ""
## " 0 - 50" "1139 (23.5)" ""
## "Viral Load at Enumeration (%)" "" "0.353"
## " < 100" "304 (22.1)" ""
## " 100 -1000" "158 (11.5)" ""
## " > 1000" "912 (66.4)" ""
## "Enumeration VL Not Done = 1 (%)" "3479 (71.7)" "0.057"
## "ART (%)" "" "NaN"
## " On ART" "0 (0.0)" ""
## " LTFU" "4853 (100.0)" ""
## " ART Naive" "0 (0.0)" ""
## "ART Naive = Subsequently Initiated (%)" "0 (NaN)" "NA"
## "Time to Initiation (median [IQR])" "NA [NA, NA]" "NA"
## "LTFU = Re-initiated (%)" "4853 (100.0)" "<0.001"
## "Time to Re-initiation (median [IQR])" "7.00 [0.00, 25.00]" "NA"
## "Enumerated in Hospital = 1 (%)" "661 (13.6)" "<0.001"
## "Current TB (PHDC) = 1 (%)" "1157 (23.8)" "0.043"
## "Previous TB (PHDC) = 1 (%)" "2292 (47.2)" "0.316"
## "Incident TB (PHDC) = 1 (%)" "998 (20.6)" "<0.001"
## "Current Cryptococcosis = 1 (%)" "77 (1.6)" "0.006"
## "Previous Cryptococcosis = 1 (%)" "46 (0.9)" "0.508"
## "Incident Cryptococcosis = 1 (%)" "102 (2.1)" "0.013"
## "Age_cat (%)" "" "0.096"
## " 25-35" "1888 (38.9)" ""
## " 18-25" "325 (6.7)" ""
## " 35-45" "1781 (36.7)" ""
## " >45" "859 (17.7)" ""
## "Stratified by LTFU"
## "" "test"
## "n" ""
## "Mortality = Died (%)" ""
## "NPR Linkage = Linked (%)" ""
## "Sex = Male (%)" ""
## "Age > 40 = 1 (%)" ""
## "Age (median [IQR])" "nonnorm"
## "Enumeration CD4 (median [IQR])" "nonnorm"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" ""
## " 51 - 100" ""
## " 0 - 50" ""
## "Viral Load at Enumeration (%)" ""
## " < 100" ""
## " 100 -1000" ""
## " > 1000" ""
## "Enumeration VL Not Done = 1 (%)" ""
## "ART (%)" ""
## " On ART" ""
## " LTFU" ""
## " ART Naive" ""
## "ART Naive = Subsequently Initiated (%)" ""
## "Time to Initiation (median [IQR])" "nonnorm"
## "LTFU = Re-initiated (%)" ""
## "Time to Re-initiation (median [IQR])" "nonnorm"
## "Enumerated in Hospital = 1 (%)" ""
## "Current TB (PHDC) = 1 (%)" ""
## "Previous TB (PHDC) = 1 (%)" ""
## "Incident TB (PHDC) = 1 (%)" ""
## "Current Cryptococcosis = 1 (%)" ""
## "Previous Cryptococcosis = 1 (%)" ""
## "Incident Cryptococcosis = 1 (%)" ""
## "Age_cat (%)" ""
## " 25-35" ""
## " 18-25" ""
## " 35-45" ""
## " >45" ""#Stratified table 3 (ART Naive/initiation status)
tab5 <- CreateTableOne(vars = myVars, strata = "ART Naive" , data = Data, factorVars = catVars)
print(tab5, nonnormal = non_normal, quote = TRUE, noSpaces = TRUE)## "Stratified by ART Naive"
## "" "Never Initiaited"
## "n" "266"
## "Mortality = Died (%)" "51 (19.2)"
## "NPR Linkage = Linked (%)" "203 (76.3)"
## "Sex = Male (%)" "138 (51.9)"
## "Age > 40 = 1 (%)" "82 (30.8)"
## "Age (median [IQR])" "35.03 [29.84, 42.39]"
## "Enumeration CD4 (median [IQR])" "103.50 [53.50, 154.00]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "134 (50.4)"
## " 51 - 100" "68 (25.6)"
## " 0 - 50" "64 (24.1)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "9 (27.3)"
## " 100 -1000" "0 (0.0)"
## " > 1000" "24 (72.7)"
## "Enumeration VL Not Done = 1 (%)" "233 (87.6)"
## "ART (%)" ""
## " On ART" "0 (0.0)"
## " LTFU" "0 (0.0)"
## " ART Naive" "266 (100.0)"
## "ART Naive = Subsequently Initiated (%)" "0 (0.0)"
## "Time to Initiation (median [IQR])" "NA [NA, NA]"
## "LTFU = Re-initiated (%)" "0 (NaN)"
## "Time to Re-initiation (median [IQR])" "NA [NA, NA]"
## "Enumerated in Hospital = 1 (%)" "75 (28.2)"
## "Current TB (PHDC) = 1 (%)" "111 (41.7)"
## "Previous TB (PHDC) = 1 (%)" "72 (27.1)"
## "Incident TB (PHDC) = 1 (%)" "41 (15.4)"
## "Current Cryptococcosis = 1 (%)" "12 (4.5)"
## "Previous Cryptococcosis = 1 (%)" "0 (0.0)"
## "Incident Cryptococcosis = 1 (%)" "3 (1.1)"
## "Age_cat (%)" ""
## " 25-35" "115 (43.2)"
## " 18-25" "17 (6.4)"
## " 35-45" "88 (33.1)"
## " >45" "46 (17.3)"
## "Stratified by ART Naive"
## "" "Subsequently Initiated" "p"
## "n" "5419" ""
## "Mortality = Died (%)" "344 (6.3)" "<0.001"
## "NPR Linkage = Linked (%)" "4525 (83.5)" "0.003"
## "Sex = Male (%)" "2599 (48.0)" "0.236"
## "Age > 40 = 1 (%)" "1606 (29.6)" "0.729"
## "Age (median [IQR])" "34.50 [29.23, 41.48]" "0.174"
## "Enumeration CD4 (median [IQR])" "108.00 [57.00, 157.00]" "0.300"
## "Enumeration CD4 Categorical (%)" "" "0.464"
## " 101 - 199" "2940 (54.3)" ""
## " 51 - 100" "1281 (23.6)" ""
## " 0 - 50" "1198 (22.1)" ""
## "Viral Load at Enumeration (%)" "" "0.034"
## " < 100" "94 (19.4)" ""
## " 100 -1000" "80 (16.5)" ""
## " > 1000" "311 (64.1)" ""
## "Enumeration VL Not Done = 1 (%)" "4934 (91.1)" "0.071"
## "ART (%)" "" "NaN"
## " On ART" "0 (0.0)" ""
## " LTFU" "0 (0.0)" ""
## " ART Naive" "5419 (100.0)" ""
## "ART Naive = Subsequently Initiated (%)" "5419 (100.0)" "<0.001"
## "Time to Initiation (median [IQR])" "10.00 [0.00, 28.00]" "NA"
## "LTFU = Re-initiated (%)" "0 (NaN)" "NA"
## "Time to Re-initiation (median [IQR])" "NA [NA, NA]" "NA"
## "Enumerated in Hospital = 1 (%)" "567 (10.5)" "<0.001"
## "Current TB (PHDC) = 1 (%)" "1644 (30.3)" "<0.001"
## "Previous TB (PHDC) = 1 (%)" "407 (7.5)" "<0.001"
## "Incident TB (PHDC) = 1 (%)" "701 (12.9)" "0.281"
## "Current Cryptococcosis = 1 (%)" "77 (1.4)" "<0.001"
## "Previous Cryptococcosis = 1 (%)" "4 (0.1)" "1.000"
## "Incident Cryptococcosis = 1 (%)" "57 (1.1)" "1.000"
## "Age_cat (%)" "" "0.309"
## " 25-35" "2298 (42.4)" ""
## " 18-25" "535 (9.9)" ""
## " 35-45" "1708 (31.5)" ""
## " >45" "878 (16.2)" ""
## "Stratified by ART Naive"
## "" "test"
## "n" ""
## "Mortality = Died (%)" ""
## "NPR Linkage = Linked (%)" ""
## "Sex = Male (%)" ""
## "Age > 40 = 1 (%)" ""
## "Age (median [IQR])" "nonnorm"
## "Enumeration CD4 (median [IQR])" "nonnorm"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" ""
## " 51 - 100" ""
## " 0 - 50" ""
## "Viral Load at Enumeration (%)" ""
## " < 100" ""
## " 100 -1000" ""
## " > 1000" ""
## "Enumeration VL Not Done = 1 (%)" ""
## "ART (%)" ""
## " On ART" ""
## " LTFU" ""
## " ART Naive" ""
## "ART Naive = Subsequently Initiated (%)" ""
## "Time to Initiation (median [IQR])" "nonnorm"
## "LTFU = Re-initiated (%)" ""
## "Time to Re-initiation (median [IQR])" "nonnorm"
## "Enumerated in Hospital = 1 (%)" ""
## "Current TB (PHDC) = 1 (%)" ""
## "Previous TB (PHDC) = 1 (%)" ""
## "Incident TB (PHDC) = 1 (%)" ""
## "Current Cryptococcosis = 1 (%)" ""
## "Previous Cryptococcosis = 1 (%)" ""
## "Incident Cryptococcosis = 1 (%)" ""
## "Age_cat (%)" ""
## " 25-35" ""
## " 18-25" ""
## " 35-45" ""
## " >45" ""#Stratified table 4 (ART)
tab6 <- CreateTableOne(vars = myVars, strata = "ART" , data = Data, factorVars = catVars)
print(tab6, nonnormal = non_normal, quote = TRUE, noSpaces = TRUE)## "Stratified by ART"
## "" "On ART"
## "n" "2379"
## "Mortality = Died (%)" "346 (14.5)"
## "NPR Linkage = Linked (%)" "1957 (82.3)"
## "Sex = Male (%)" "1003 (42.2)"
## "Age > 40 = 1 (%)" "1062 (44.6)"
## "Age (median [IQR])" "38.56 [32.65, 45.91]"
## "Enumeration CD4 (median [IQR])" "126.00 [74.00, 166.00]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "1510 (63.5)"
## " 51 - 100" "489 (20.6)"
## " 0 - 50" "380 (16.0)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "880 (44.0)"
## " 100 -1000" "174 (8.7)"
## " > 1000" "948 (47.4)"
## "Enumeration VL Not Done = 1 (%)" "377 (15.8)"
## "ART (%)" ""
## " On ART" "2379 (100.0)"
## " LTFU" "0 (0.0)"
## " ART Naive" "0 (0.0)"
## "ART Naive = Subsequently Initiated (%)" "0 (NaN)"
## "Time to Initiation (median [IQR])" "NA [NA, NA]"
## "LTFU = Re-initiated (%)" "0 (NaN)"
## "Time to Re-initiation (median [IQR])" "NA [NA, NA]"
## "Enumerated in Hospital = 1 (%)" "566 (23.8)"
## "Current TB (PHDC) = 1 (%)" "367 (15.4)"
## "Previous TB (PHDC) = 1 (%)" "1410 (59.3)"
## "Incident TB (PHDC) = 1 (%)" "437 (18.4)"
## "Current Cryptococcosis = 1 (%)" "44 (1.8)"
## "Previous Cryptococcosis = 1 (%)" "63 (2.6)"
## "Incident Cryptococcosis = 1 (%)" "49 (2.1)"
## "Age_cat (%)" ""
## " 25-35" "713 (30.0)"
## " 18-25" "123 (5.2)"
## " 35-45" "890 (37.4)"
## " >45" "653 (27.4)"
## "Stratified by ART"
## "" "LTFU"
## "n" "5280"
## "Mortality = Died (%)" "608 (11.5)"
## "NPR Linkage = Linked (%)" "4467 (84.6)"
## "Sex = Male (%)" "1988 (37.7)"
## "Age > 40 = 1 (%)" "1753 (33.2)"
## "Age (median [IQR])" "36.02 [30.74, 42.43]"
## "Enumeration CD4 (median [IQR])" "106.00 [52.00, 155.00]"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" "2784 (52.7)"
## " 51 - 100" "1235 (23.4)"
## " 0 - 50" "1261 (23.9)"
## "Viral Load at Enumeration (%)" ""
## " < 100" "328 (21.7)"
## " 100 -1000" "177 (11.7)"
## " > 1000" "1009 (66.6)"
## "Enumeration VL Not Done = 1 (%)" "3766 (71.3)"
## "ART (%)" ""
## " On ART" "0 (0.0)"
## " LTFU" "5280 (100.0)"
## " ART Naive" "0 (0.0)"
## "ART Naive = Subsequently Initiated (%)" "0 (NaN)"
## "Time to Initiation (median [IQR])" "NA [NA, NA]"
## "LTFU = Re-initiated (%)" "4853 (91.9)"
## "Time to Re-initiation (median [IQR])" "7.00 [0.00, 25.00]"
## "Enumerated in Hospital = 1 (%)" "855 (16.2)"
## "Current TB (PHDC) = 1 (%)" "1278 (24.2)"
## "Previous TB (PHDC) = 1 (%)" "2505 (47.4)"
## "Incident TB (PHDC) = 1 (%)" "1033 (19.6)"
## "Current Cryptococcosis = 1 (%)" "92 (1.7)"
## "Previous Cryptococcosis = 1 (%)" "52 (1.0)"
## "Incident Cryptococcosis = 1 (%)" "103 (2.0)"
## "Age_cat (%)" ""
## " 25-35" "2028 (38.4)"
## " 18-25" "356 (6.7)"
## " 35-45" "1951 (37.0)"
## " >45" "945 (17.9)"
## "Stratified by ART"
## "" "ART Naive" "p"
## "n" "5685" ""
## "Mortality = Died (%)" "395 (6.9)" "<0.001"
## "NPR Linkage = Linked (%)" "4728 (83.2)" "0.021"
## "Sex = Male (%)" "2737 (48.1)" "<0.001"
## "Age > 40 = 1 (%)" "1688 (29.7)" "<0.001"
## "Age (median [IQR])" "34.53 [29.29, 41.53]" "<0.001"
## "Enumeration CD4 (median [IQR])" "108.00 [56.00, 157.00]" "<0.001"
## "Enumeration CD4 Categorical (%)" "" "<0.001"
## " 101 - 199" "3074 (54.1)" ""
## " 51 - 100" "1349 (23.7)" ""
## " 0 - 50" "1262 (22.2)" ""
## "Viral Load at Enumeration (%)" "" "<0.001"
## " < 100" "103 (19.9)" ""
## " 100 -1000" "80 (15.4)" ""
## " > 1000" "335 (64.7)" ""
## "Enumeration VL Not Done = 1 (%)" "5167 (90.9)" "<0.001"
## "ART (%)" "" "<0.001"
## " On ART" "0 (0.0)" ""
## " LTFU" "0 (0.0)" ""
## " ART Naive" "5685 (100.0)" ""
## "ART Naive = Subsequently Initiated (%)" "5419 (95.3)" "NaN"
## "Time to Initiation (median [IQR])" "10.00 [0.00, 28.00]" "NA"
## "LTFU = Re-initiated (%)" "0 (NaN)" "NaN"
## "Time to Re-initiation (median [IQR])" "NA [NA, NA]" "NA"
## "Enumerated in Hospital = 1 (%)" "642 (11.3)" "<0.001"
## "Current TB (PHDC) = 1 (%)" "1755 (30.9)" "<0.001"
## "Previous TB (PHDC) = 1 (%)" "479 (8.4)" "<0.001"
## "Incident TB (PHDC) = 1 (%)" "742 (13.1)" "<0.001"
## "Current Cryptococcosis = 1 (%)" "89 (1.6)" "0.612"
## "Previous Cryptococcosis = 1 (%)" "4 (0.1)" "<0.001"
## "Incident Cryptococcosis = 1 (%)" "60 (1.1)" "<0.001"
## "Age_cat (%)" "" "<0.001"
## " 25-35" "2413 (42.4)" ""
## " 18-25" "552 (9.7)" ""
## " 35-45" "1796 (31.6)" ""
## " >45" "924 (16.3)" ""
## "Stratified by ART"
## "" "test"
## "n" ""
## "Mortality = Died (%)" ""
## "NPR Linkage = Linked (%)" ""
## "Sex = Male (%)" ""
## "Age > 40 = 1 (%)" ""
## "Age (median [IQR])" "nonnorm"
## "Enumeration CD4 (median [IQR])" "nonnorm"
## "Enumeration CD4 Categorical (%)" ""
## " 101 - 199" ""
## " 51 - 100" ""
## " 0 - 50" ""
## "Viral Load at Enumeration (%)" ""
## " < 100" ""
## " 100 -1000" ""
## " > 1000" ""
## "Enumeration VL Not Done = 1 (%)" ""
## "ART (%)" ""
## " On ART" ""
## " LTFU" ""
## " ART Naive" ""
## "ART Naive = Subsequently Initiated (%)" ""
## "Time to Initiation (median [IQR])" "nonnorm"
## "LTFU = Re-initiated (%)" ""
## "Time to Re-initiation (median [IQR])" "nonnorm"
## "Enumerated in Hospital = 1 (%)" ""
## "Current TB (PHDC) = 1 (%)" ""
## "Previous TB (PHDC) = 1 (%)" ""
## "Incident TB (PHDC) = 1 (%)" ""
## "Current Cryptococcosis = 1 (%)" ""
## "Previous Cryptococcosis = 1 (%)" ""
## "Incident Cryptococcosis = 1 (%)" ""
## "Age_cat (%)" ""
## " 25-35" ""
## " 18-25" ""
## " 35-45" ""
## " >45" ""Survival anaylsis
### Overall ART KM
ggsurvplot(
fit = survfit(Surv(Time, Censored) ~ ART, data = Linked_Data),
risk.table = TRUE,
cumevents = TRUE,
conf.int = T,
pval = TRUE,
palette = "npg",
linetype = "solid",
risk.table.height = 0.2,
cumevents.height = 0.2,
xlim = c(0,1000),
xlab = "Days",
break.time.by = 90,
ylim = c(0, 1),
ylab = "Survival",
legend = "top",
title = "Time to death among linked patients stratifed by ART exposure at enumeration")
### Zoomed in Overall ART KM
ggsurvplot(
fit = survfit(Surv(Time, Censored) ~ ART, data = Linked_Data),
legend = "none",
risk.table = F,
cumevents = F,
conf.int = T,
pval = F,
palette = "npg",
linetype = "solid",
xlim = c(0,1000),
break.time.by = 180,
ylim = c(0.8, 1),
xlab = "",
ylab = "")
### VL KM
### On ART by VL KM
L_Data <- subset(Data, `NPR linkage` = "Linked")
On_ART <- subset(L_Data, ART == "On ART")
On_ART$VL <- On_ART$`Viral Load at Enumeration`
summary(On_ART$VL)## < 100 100 -1000 > 1000 NA's
## 880 174 948 377ggsurvplot(
fit = survfit(Surv(Time, Censored) ~ VL, data = On_ART),
risk.table = TRUE,
cumevents = TRUE,
conf.int = T,
pval = TRUE,
palette = "npg",
linetype = "solid",
risk.table.height = 0.2,
cumevents.height = 0.2,
xlim = c(0,1000),
xlab = "Days",
break.time.by = 90,
ylim = c(0, 1),
ylab = "Survival",
legend = "top",
title = "Time to death among patients on ART at enumeration stratified by enumeration viral load")
# VL zoomed in
ggsurvplot(
fit = survfit(Surv(Time, Censored) ~ VL, data = On_ART),
legend = "none",
risk.table = F,
cumevents = F,
conf.int = T,
pval = F,
palette = "npg",
linetype = "solid",
xlim = c(0,1000),
break.time.by = 180,
ylim = c(0.8, 1),
xlab = "",
ylab = "")
Modelling
Pre-process data again
## Processing data
Cohort$VL_E_Cat[is.na(Cohort$VL_E_Cat)] = 3
Cohort$Enumeration_CD4 <- as.numeric(Cohort$Enumeration_CD4)
Data <- select(Cohort, study_id, Censored, Time, Enumeration_CD4, VL_E_Cat, E_CD4_Cat, Hospital_Enumerations,
Incident_Crypto, Prevalent_Crypto, Previous_Crypto,
ART_Exp_C, Subs_Initiated, Time_To_First_ART, Reinitiated, Time_Reinitiation
, Sex_B, Age, Age_Cat, Linked, Current_Diabetes, Current_Hypertension,
Current_CKD, Current_TB_PHDC, Previous_TB_PHDC, Incident_TB_PHDC)
summary(Data)## study_id Censored Time Enumeration_CD4
## Min. :1.001e+09 Min. :0.0000 Min. : 0.0 Min. : 0.0
## 1st Qu.:3.194e+09 1st Qu.:0.0000 1st Qu.: 489.8 1st Qu.: 57.0
## Median :5.430e+09 Median :0.0000 Median : 895.5 Median :111.0
## Mean :5.441e+09 Mean :0.1011 Mean : 856.1 Mean :106.9
## 3rd Qu.:7.664e+09 3rd Qu.:0.0000 3rd Qu.:1272.0 3rd Qu.:158.0
## Max. :9.998e+09 Max. :1.0000 Max. :1620.0 Max. :199.0
##
## VL_E_Cat E_CD4_Cat Hospital_Enumerations Incident_Crypto
## Min. :0.000 Min. :0.0000 Min. :0.0000 Min. :0.00000
## 1st Qu.:2.000 1st Qu.:0.0000 1st Qu.:0.0000 1st Qu.:0.00000
## Median :3.000 Median :0.0000 Median :0.0000 Median :0.00000
## Mean :2.469 Mean :0.6654 Mean :0.1546 Mean :0.01589
## 3rd Qu.:3.000 3rd Qu.:1.0000 3rd Qu.:0.0000 3rd Qu.:0.00000
## Max. :3.000 Max. :2.0000 Max. :1.0000 Max. :1.00000
##
## Prevalent_Crypto Previous_Crypto ART_Exp_C Subs_Initiated
## Min. :0.00000 Min. :0.000000 Length:13344 Min. :0.000
## 1st Qu.:0.00000 1st Qu.:0.000000 Class :character 1st Qu.:1.000
## Median :0.00000 Median :0.000000 Mode :character Median :1.000
## Mean :0.01686 Mean :0.008918 Mean :0.953
## 3rd Qu.:0.00000 3rd Qu.:0.000000 3rd Qu.:1.000
## Max. :1.00000 Max. :1.000000 Max. :1.000
## NA's :7659
## Time_To_First_ART Reinitiated Time_Reinitiation Sex_B
## Min. : 0.00 Min. :0.000 Min. : 0.0 Min. :0.0000
## 1st Qu.: 0.00 1st Qu.:1.000 1st Qu.: 0.0 1st Qu.:0.0000
## Median : 10.00 Median :1.000 Median : 7.0 Median :0.0000
## Mean : 49.62 Mean :0.919 Mean : 39.4 Mean :0.4293
## 3rd Qu.: 28.00 3rd Qu.:1.000 3rd Qu.: 25.0 3rd Qu.:1.0000
## Max. :1503.00 Max. :1.000 Max. :1287.0 Max. :1.0000
## NA's :7925 NA's :8064 NA's :8491
## Age Age_Cat Linked Current_Diabetes
## Min. :18.01 Min. :0.0000 Min. :0.0000 Min. :0.00000
## 1st Qu.:30.34 1st Qu.:0.0000 1st Qu.:1.0000 1st Qu.:0.00000
## Median :35.78 Median :0.0000 Median :1.0000 Median :0.00000
## Mean :37.08 Mean :0.3375 Mean :0.8357 Mean :0.02608
## 3rd Qu.:42.70 3rd Qu.:1.0000 3rd Qu.:1.0000 3rd Qu.:0.00000
## Max. :89.25 Max. :1.0000 Max. :1.0000 Max. :1.00000
##
## Current_Hypertension Current_CKD Current_TB_PHDC Previous_TB_PHDC
## Min. :0.00000 Min. :0.00000 Min. :0.0000 Min. :0.0000
## 1st Qu.:0.00000 1st Qu.:0.00000 1st Qu.:0.0000 1st Qu.:0.0000
## Median :0.00000 Median :0.00000 Median :0.0000 Median :0.0000
## Mean :0.08933 Mean :0.01963 Mean :0.2548 Mean :0.3293
## 3rd Qu.:0.00000 3rd Qu.:0.00000 3rd Qu.:1.0000 3rd Qu.:1.0000
## Max. :1.00000 Max. :1.00000 Max. :1.0000 Max. :1.0000
##
## Incident_TB_PHDC
## Min. :0.0000
## 1st Qu.:0.0000
## Median :0.0000
## Mean :0.1658
## 3rd Qu.:0.0000
## Max. :1.0000
## table(Data$ART_Exp_C)##
## LTFU Naive On ART
## 5280 5685 2379table(Data$Age_Cat)##
## 0 1
## 8841 4503Data <- within(Data, {
`Viral Load at Enumeration` <- factor(VL_E_Cat, labels = c("< 100", "100 -1000", "> 1000", "None"))
ART <- factor(ART_Exp_C, levels = c("On ART","LTFU","Naive"), labels = c("On ART","LTFU","ART Naive"))
Disengaged <- factor(Reinitiated, labels = c("Never Re-initiated", "Re-initiated"))
`ART Naive` <- factor(Subs_Initiated, labels = c("Never Initiaited", "Subsequently Initiated"))
`Enumeration CD4` <- Enumeration_CD4
`Enumeration CD4 Categorical` <- factor(E_CD4_Cat, labels = c("101 - 199", "51 - 100", "0 - 50"))
`Time in Cohort` <- Time
`Time to Initiation` <- Time_To_First_ART
`Time to Re-initiation` <- Time_Reinitiation
Mortality <- factor(Censored, labels = c("Survived", "Died"))
Sex <- factor(Sex_B, labels = c("Female", "Male"))
`Sex (Male)` <- Sex_B
`Age` <- Age
`Incident Cryptococcosis` <- Incident_Crypto
`Current Cryptococcosis` <- Prevalent_Crypto
`Previous Cryptococcosis` <- Previous_Crypto
`Enumerated in Hospital` <- Hospital_Enumerations
`NPR Linkage` <- factor(Linked, labels = c("Unlinked", "Linked"))
`Diabetic` <- Current_Diabetes
`Hypertension` <- Current_Hypertension
`Chronic Kidney Disease` <- Current_CKD
`Current TB (PHDC)` <- Current_TB_PHDC
`Previous TB (PHDC)` <- Previous_TB_PHDC
`Incident TB (PHDC)` <- Incident_TB_PHDC
})
Data <- select(Data, study_id, `Time`, `Viral Load at Enumeration`,
`Enumeration CD4`, `Enumerated in Hospital`, `Enumeration CD4 Categorical`,
`NPR Linkage`, `Mortality`,
`Sex`, `Sex (Male)`, `Age`,
`Incident TB (PHDC)`, `Previous TB (PHDC)`, `Current TB (PHDC)`,
`Previous Cryptococcosis`, `Current Cryptococcosis`, `Incident Cryptococcosis`,
`Chronic Kidney Disease`, `Hypertension`, `Diabetic`,
`ART Naive`, `Disengaged`, `ART`,
`Time to Initiation`, `Time to Re-initiation`, Censored, VL_E_Cat)
Linked_Data <- subset(Data, `NPR Linkage` == "Linked")
Linked_Data$AgeCat_4<-cut(Linked_Data$Age, c(0,25,35,45,100))
summary(Linked_Data$AgeCat_4)## (0,25] (25,35] (35,45] (45,100]
## 895 4401 3831 2025Linked_Data$Age_Cat_4 <- NA
Linked_Data$Age_Cat_4[Linked_Data$AgeCat_4=="(0,25]"] <- 0
Linked_Data$Age_Cat_4[Linked_Data$AgeCat_4=="(25,35]"] <- 1
Linked_Data$Age_Cat_4[Linked_Data$AgeCat_4=="(35,45]"] <- 2
Linked_Data$Age_Cat_4[Linked_Data$AgeCat_4=="(45,100]"] <- 3
table(Linked_Data$Age_Cat_4)##
## 0 1 2 3
## 895 4401 3831 2025Model_Data <- Linked_Data[, c("study_id", "Censored", "Time", "Age_Cat_4", "ART",
"Sex (Male)", "Enumeration CD4 Categorical","Previous TB (PHDC)",
"Incident TB (PHDC)" , "Current TB (PHDC)",
"Previous Cryptococcosis", "Viral Load at Enumeration", "Enumerated in Hospital")]
names(Model_Data)[4] <- "Age"
Model_Data <- within(Model_Data, {
Age <- factor(Age, labels = c("18-25", "25-35", "35-45", ">45"))})
Model_Data$Age = relevel(Model_Data$Age, ref = "25-35")
Model_Data$Time <- Model_Data$Time + 0.5
names(Model_Data)[7] <- "CD4"
names(Model_Data)[8] <- "Previous TB"
names(Model_Data)[9] <- "Incident TB"
names(Model_Data)[10] <- "Current TB"
head(Model_Data)## # A tibble: 6 × 13
## study_id Censored Time Age ART `Sex (Male)` CD4 `Previous TB`
## <dbl> <dbl> <dbl> <fct> <fct> <dbl> <fct> <dbl>
## 1 1001292400 0 874. 35-45 ART Naive 1 101 - 199 0
## 2 1001394208 0 1296. 25-35 ART Naive 1 101 - 199 0
## 3 1001490301 0 1462. 35-45 On ART 0 0 - 50 1
## 4 1001698318 0 1458. 18-25 On ART 1 101 - 199 0
## 5 1002697417 0 1472. 35-45 On ART 1 101 - 199 0
## 6 1002894013 0 1546. 25-35 On ART 0 101 - 199 0
## # ℹ 5 more variables: `Incident TB` <dbl>, `Current TB` <dbl>,
## # `Previous Cryptococcosis` <dbl>, `Viral Load at Enumeration` <fct>,
## # `Enumerated in Hospital` <dbl>CPH Model 1
This model includes the variables: ART CD4 cat at enumeration Sex Age Previous TB Previous Crypto
## Call:
## coxph(formula = Surv(Time, Censored) ~ ART + `Sex (Male)` + Age +
## CD4 + `Previous TB` + `Previous Cryptococcosis`, data = Model_Data)
##
## n= 11152, number of events= 1349
##
## coef exp(coef) se(coef) z Pr(>|z|)
## ARTLTFU -0.19589 0.82210 0.06875 -2.849 0.00438 **
## ARTART Naive -0.51801 0.59570 0.08381 -6.181 6.37e-10 ***
## `Sex (Male)` 0.03560 1.03625 0.05734 0.621 0.53464
## Age18-25 0.17795 1.19476 0.11642 1.528 0.12640
## Age35-45 0.17966 1.19681 0.06784 2.648 0.00809 **
## Age>45 0.56229 1.75468 0.07445 7.553 4.26e-14 ***
## CD451 - 100 0.48027 1.61651 0.07156 6.711 1.93e-11 ***
## CD40 - 50 1.06478 2.90020 0.06364 16.730 < 2e-16 ***
## `Previous TB` 0.38818 1.47430 0.06364 6.100 1.06e-09 ***
## `Previous Cryptococcosis` -0.12524 0.88228 0.24566 -0.510 0.61018
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## ARTLTFU 0.8221 1.2164 0.7185 0.9407
## ARTART Naive 0.5957 1.6787 0.5055 0.7021
## `Sex (Male)` 1.0362 0.9650 0.9261 1.1595
## Age18-25 1.1948 0.8370 0.9510 1.5010
## Age35-45 1.1968 0.8356 1.0478 1.3670
## Age>45 1.7547 0.5699 1.5165 2.0303
## CD451 - 100 1.6165 0.6186 1.4050 1.8599
## CD40 - 50 2.9002 0.3448 2.5601 3.2855
## `Previous TB` 1.4743 0.6783 1.3014 1.6701
## `Previous Cryptococcosis` 0.8823 1.1334 0.5451 1.4280
##
## Concordance= 0.689 (se = 0.007 )
## Likelihood ratio test= 516.9 on 10 df, p=<2e-16
## Wald test = 527.7 on 10 df, p=<2e-16
## Score (logrank) test = 562.9 on 10 df, p=<2e-16It is important to note that the diagonstics of model 1 raise concerns about the PH assumption:
The statistically significant variables are age, CD4 count and ART status. A further analysis of these features is conducted after the uni-variate modelling. If one looks at the plots they highlight most of the hazzard diversion happen around the first +- year (with the worst in the first +- 90 days). Given p-values are not be the best way to evaluate the PH assumption given the large sample size the PH assumption has been evaluated for these features using KM plots, log-log curves and schoenfeld residuals. In this process it is noted that ART status is the most important violation of the PH assumption thus to better unpack this, stratifed models for each ART status are presented.
Stratifed CPH models for each ART category
On_ART <- subset(Model_Data, ART == "On ART")
Disengaged <- subset(Model_Data, ART == "LTFU")
Naive <- subset(Model_Data, ART == "ART Naive")
Surv_object_On_ART <- Surv(time = On_ART$Time, event = On_ART$Censored)
Surv_object_Disengaged <- Surv(time = Disengaged$Time, event = Disengaged$Censored)
Surv_object_Naive <- Surv(time = Naive$Time, event = Naive$Censored)
# Stratified Surv models
surv_ART1 <- coxph(Surv_object_On_ART ~ `Sex (Male)` + `Age` + `CD4` +
`Previous TB` + `Previous Cryptococcosis`,data = On_ART)
summary(surv_ART1)## Call:
## coxph(formula = Surv_object_On_ART ~ `Sex (Male)` + Age + CD4 +
## `Previous TB` + `Previous Cryptococcosis`, data = On_ART)
##
## n= 1957, number of events= 346
##
## coef exp(coef) se(coef) z Pr(>|z|)
## `Sex (Male)` -0.052477 0.948876 0.113747 -0.461 0.64455
## Age18-25 -0.286261 0.751067 0.294756 -0.971 0.33146
## Age35-45 0.003349 1.003355 0.135910 0.025 0.98034
## Age>45 0.392014 1.479959 0.143009 2.741 0.00612 **
## CD451 - 100 0.528239 1.695943 0.133969 3.943 8.05e-05 ***
## CD40 - 50 0.953167 2.593911 0.129839 7.341 2.12e-13 ***
## `Previous TB` 0.369100 1.446432 0.120020 3.075 0.00210 **
## `Previous Cryptococcosis` -0.199629 0.819035 0.339536 -0.588 0.55657
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## `Sex (Male)` 0.9489 1.0539 0.7593 1.186
## Age18-25 0.7511 1.3314 0.4215 1.338
## Age35-45 1.0034 0.9967 0.7687 1.310
## Age>45 1.4800 0.6757 1.1182 1.959
## CD451 - 100 1.6959 0.5896 1.3043 2.205
## CD40 - 50 2.5939 0.3855 2.0111 3.346
## `Previous TB` 1.4464 0.6914 1.1432 1.830
## `Previous Cryptococcosis` 0.8190 1.2209 0.4210 1.593
##
## Concordance= 0.646 (se = 0.015 )
## Likelihood ratio test= 76.35 on 8 df, p=3e-13
## Wald test = 78.46 on 8 df, p=1e-13
## Score (logrank) test = 82.32 on 8 df, p=2e-14surv_ART2 <- coxph(Surv_object_On_ART ~ `Sex (Male)` + `Age` + `Viral Load at Enumeration` +
`Previous TB` + `Previous Cryptococcosis`,data = On_ART)
summary(surv_ART2)## Call:
## coxph(formula = Surv_object_On_ART ~ `Sex (Male)` + Age + `Viral Load at Enumeration` +
## `Previous TB` + `Previous Cryptococcosis`, data = On_ART)
##
## n= 1957, number of events= 346
##
## coef exp(coef) se(coef) z
## `Sex (Male)` 0.019700 1.019895 0.113718 0.173
## Age18-25 -0.252933 0.776520 0.294940 -0.858
## Age35-45 -0.004276 0.995733 0.136112 -0.031
## Age>45 0.342815 1.408908 0.142762 2.401
## `Viral Load at Enumeration`100 -1000 0.038080 1.038815 0.243825 0.156
## `Viral Load at Enumeration`> 1000 0.408996 1.505305 0.125272 3.265
## `Viral Load at Enumeration`None 0.225725 1.253231 0.171849 1.314
## `Previous TB` 0.407826 1.503546 0.119965 3.400
## `Previous Cryptococcosis` -0.143137 0.866636 0.339208 -0.422
## Pr(>|z|)
## `Sex (Male)` 0.862467
## Age18-25 0.391128
## Age35-45 0.974938
## Age>45 0.016337 *
## `Viral Load at Enumeration`100 -1000 0.875892
## `Viral Load at Enumeration`> 1000 0.001095 **
## `Viral Load at Enumeration`None 0.189011
## `Previous TB` 0.000675 ***
## `Previous Cryptococcosis` 0.673045
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## `Sex (Male)` 1.0199 0.9805 0.8161 1.275
## Age18-25 0.7765 1.2878 0.4356 1.384
## Age35-45 0.9957 1.0043 0.7626 1.300
## Age>45 1.4089 0.7098 1.0650 1.864
## `Viral Load at Enumeration`100 -1000 1.0388 0.9626 0.6442 1.675
## `Viral Load at Enumeration`> 1000 1.5053 0.6643 1.1776 1.924
## `Viral Load at Enumeration`None 1.2532 0.7979 0.8949 1.755
## `Previous TB` 1.5035 0.6651 1.1885 1.902
## `Previous Cryptococcosis` 0.8666 1.1539 0.4458 1.685
##
## Concordance= 0.586 (se = 0.016 )
## Likelihood ratio test= 35.04 on 9 df, p=6e-05
## Wald test = 34.35 on 9 df, p=8e-05
## Score (logrank) test = 34.8 on 9 df, p=6e-05surv_Dis <- coxph(Surv_object_Disengaged ~ `Sex (Male)` + `Age` + `CD4` +
`Previous TB` + `Previous Cryptococcosis`,data = Disengaged)
summary(surv_Dis)## Call:
## coxph(formula = Surv_object_Disengaged ~ `Sex (Male)` + Age +
## CD4 + `Previous TB` + `Previous Cryptococcosis`, data = Disengaged)
##
## n= 4467, number of events= 608
##
## coef exp(coef) se(coef) z Pr(>|z|)
## `Sex (Male)` -0.02809 0.97230 0.08856 -0.317 0.75107
## Age18-25 0.43421 1.54375 0.16148 2.689 0.00717 **
## Age35-45 0.09664 1.10147 0.10069 0.960 0.33713
## Age>45 0.52086 1.68348 0.11541 4.513 6.39e-06 ***
## CD451 - 100 0.53898 1.71426 0.11172 4.824 1.41e-06 ***
## CD40 - 50 1.22582 3.40697 0.09605 12.763 < 2e-16 ***
## `Previous TB` 0.40420 1.49811 0.08575 4.714 2.43e-06 ***
## `Previous Cryptococcosis` 0.02175 1.02199 0.35729 0.061 0.95146
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## `Sex (Male)` 0.9723 1.0285 0.8174 1.157
## Age18-25 1.5437 0.6478 1.1249 2.118
## Age35-45 1.1015 0.9079 0.9042 1.342
## Age>45 1.6835 0.5940 1.3427 2.111
## CD451 - 100 1.7143 0.5833 1.3771 2.134
## CD40 - 50 3.4070 0.2935 2.8224 4.113
## `Previous TB` 1.4981 0.6675 1.2663 1.772
## `Previous Cryptococcosis` 1.0220 0.9785 0.5074 2.059
##
## Concordance= 0.688 (se = 0.011 )
## Likelihood ratio test= 235.8 on 8 df, p=<2e-16
## Wald test = 237.7 on 8 df, p=<2e-16
## Score (logrank) test = 260.7 on 8 df, p=<2e-16surv_Naive <- coxph(Surv_object_Naive ~ `Sex (Male)` + `Age` + `CD4` +
`Previous TB`,data = Naive)
summary(surv_Naive)## Call:
## coxph(formula = Surv_object_Naive ~ `Sex (Male)` + Age + CD4 +
## `Previous TB`, data = Naive)
##
## n= 4728, number of events= 395
##
## coef exp(coef) se(coef) z Pr(>|z|)
## `Sex (Male)` 0.2298 1.2583 0.1037 2.215 0.026750 *
## Age18-25 0.1452 1.1563 0.2083 0.697 0.485841
## Age35-45 0.4542 1.5749 0.1246 3.647 0.000266 ***
## Age>45 0.8127 2.2539 0.1356 5.995 2.04e-09 ***
## CD451 - 100 0.3755 1.4557 0.1304 2.879 0.003994 **
## CD40 - 50 0.8974 2.4532 0.1173 7.650 2.00e-14 ***
## `Previous TB` 0.4299 1.5370 0.1530 2.810 0.004947 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## exp(coef) exp(-coef) lower .95 upper .95
## `Sex (Male)` 1.258 0.7947 1.0268 1.542
## Age18-25 1.156 0.8648 0.7686 1.739
## Age35-45 1.575 0.6350 1.2338 2.010
## Age>45 2.254 0.4437 1.7280 2.940
## CD451 - 100 1.456 0.6870 1.1273 1.880
## CD40 - 50 2.453 0.4076 1.9493 3.087
## `Previous TB` 1.537 0.6506 1.1389 2.074
##
## Concordance= 0.676 (se = 0.014 )
## Likelihood ratio test= 122.2 on 7 df, p=<2e-16
## Wald test = 127.9 on 7 df, p=<2e-16
## Score (logrank) test = 135.3 on 7 df, p=<2e-16Univariate modelling
## Univariate associations
univ_variables <- c("Age", "ART",
"Sex (Male)", "CD4","Previous TB",
"Incident TB" , "Current TB",
"Previous Cryptococcosis", "Viral Load at Enumeration", "Enumerated in Hospital")
univ_variables_strat <- c("Age", "Sex (Male)", "CD4","Previous TB",
"Incident TB" , "Current TB",
"Previous Cryptococcosis", "Viral Load at Enumeration", "Enumerated in Hospital")
univ_variables_strat2 <- c("Age", "Sex (Male)", "CD4","Previous TB",
"Incident TB" , "Current TB", "Enumerated in Hospital")
UV_all <- Model_Data %>%
dplyr::select(univ_variables, Censored) %>% # select variables to model
tbl_uvregression( # produce univariate table
method = glm, ## define regression
y = Censored, ## define outcome
method.args = list(family = binomial(log)), ## define type of glm
exponentiate = TRUE ## exponentiate to produce RR
)
UV_all <- UV_all %>%
modify_table_body(
~ .x %>% select(-p.value)
)
UV_ART <- On_ART %>%
dplyr::select(univ_variables_strat, Censored) %>%
tbl_uvregression(
method = glm,
y = Censored,
method.args = list(family = binomial(log)),
exponentiate = TRUE
)
UV_ART <- UV_ART %>%
modify_table_body(
~ .x %>% select(-p.value)
)
UV_Dis <- Disengaged %>%
dplyr::select(univ_variables_strat, Censored) %>%
tbl_uvregression(
method = glm,
y = Censored,
method.args = list(family = binomial(log)),
exponentiate = TRUE
)
UV_Dis <- UV_Dis %>%
modify_table_body(
~ .x %>% select(-p.value)
)
UV_Naive <- Naive %>%
dplyr::select(univ_variables_strat2, Censored) %>%
tbl_uvregression(
method = glm,
y = Censored,
method.args = list(family = binomial(log)),
exponentiate = TRUE
)
UV_Naive <- UV_Naive %>%
modify_table_body(
~ .x %>% select(-p.value)
)CPH univariate
## Univariate associations using Cox models
library(survival)
library(dplyr)
library(gtsummary)
univ_variables <- c("Age", "ART", "Sex (Male)", "CD4", "Previous TB", "Incident TB", "Current TB", "Previous Cryptococcosis", "Viral Load at Enumeration", "Enumerated in Hospital")
univ_variables_strat <- c("Age", "Sex (Male)", "CD4", "Previous TB", "Incident TB", "Current TB", "Previous Cryptococcosis", "Viral Load at Enumeration", "Enumerated in Hospital")
univ_variables_strat2 <- c("Age", "Sex (Male)", "CD4", "Previous TB", "Incident TB", "Current TB", "Enumerated in Hospital")
# Create survival objects
Model_Data <- Model_Data #%>% mutate(Time = ..., Censored = ...) # Ensure Time and Censored columns are correctly set
Surv_object_Model_Data <- Surv(time = Model_Data$Time, event = Model_Data$Censored)
On_ART <- subset(Model_Data, ART == "On ART")
Disengaged <- subset(Model_Data, ART == "LTFU")
Naive <- subset(Model_Data, ART == "ART Naive")
Surv_object_On_ART <- Surv(time = On_ART$Time, event = On_ART$Censored)
Surv_object_Disengaged <- Surv(time = Disengaged$Time, event = Disengaged$Censored)
Surv_object_Naive <- Surv(time = Naive$Time, event = Naive$Censored)
# Univariate Cox models for all data
UV_all <- Model_Data %>%
dplyr::select(all_of(univ_variables), Time, Censored) %>% # select variables to model
tbl_uvregression( # produce univariate table
method = coxph, ## define regression
y = Surv(Time, Censored), ## define outcome
exponentiate = TRUE ## exponentiate to produce HR
)
UV_all <- UV_all %>%
modify_table_body(
~ .x %>% select(-p.value)
)
# Univariate Cox models for On ART group
UV_ART <- On_ART %>%
dplyr::select(all_of(univ_variables_strat), Time, Censored) %>%
tbl_uvregression(
method = coxph,
y = Surv(Time, Censored),
exponentiate = TRUE
)
UV_ART <- UV_ART %>%
modify_table_body(
~ .x %>% select(-p.value)
)
# Univariate Cox models for Disengaged group
UV_Dis <- Disengaged %>%
dplyr::select(all_of(univ_variables_strat), Time, Censored) %>%
tbl_uvregression(
method = coxph,
y = Surv(Time, Censored),
exponentiate = TRUE
)
UV_Dis <- UV_Dis %>%
modify_table_body(
~ .x %>% select(-p.value)
)
# Univariate Cox models for Naive group
UV_Naive <- Naive %>%
dplyr::select(all_of(univ_variables_strat2), Time, Censored) %>%
tbl_uvregression(
method = coxph,
y = Surv(Time, Censored),
exponentiate = TRUE
)
UV_Naive <- UV_Naive %>%
modify_table_body(
~ .x %>% select(-p.value)
)
# Display tables
UV_all| Characteristic | N | HR | 95% CI |
|---|---|---|---|
| Age | 11,152 | ||
|     25-35 | — | — | |
| Â Â Â Â 18-25 | 1.03 | 0.82, 1.29 | |
| Â Â Â Â 35-45 | 1.36 | 1.19, 1.55 | |
| Â Â Â Â >45 | 1.94 | 1.68, 2.23 | |
| ART | 11,152 | ||
|     On ART | — | — | |
| Â Â Â Â LTFU | 0.84 | 0.74, 0.96 | |
| Â Â Â Â ART Naive | 0.49 | 0.43, 0.57 | |
| Sex (Male) | 11,152 | 1.22 | 1.09, 1.36 |
| CD4 | 11,152 | ||
|     101 - 199 | — | — | |
| Â Â Â Â 51 - 100 | 1.60 | 1.39, 1.84 | |
| Â Â Â Â 0 - 50 | 2.94 | 2.60, 3.33 | |
| Previous TB | 11,152 | 2.00 | 1.80, 2.23 |
| Incident TB | 11,152 | 1.34 | 1.19, 1.52 |
| Current TB | 11,152 | 1.88 | 1.68, 2.10 |
| Previous Cryptococcosis | 11,152 | 1.40 | 0.87, 2.25 |
| Viral Load at Enumeration | 11,152 | ||
|     < 100 | — | — | |
| Â Â Â Â 100 -1000 | 0.97 | 0.70, 1.35 | |
| Â Â Â Â > 1000 | 1.53 | 1.27, 1.85 | |
| Â Â Â Â None | 0.78 | 0.65, 0.93 | |
| Enumerated in Hospital | 11,152 | 3.11 | 2.77, 3.49 |
| Abbreviations: CI = Confidence Interval, HR = Hazard Ratio | |||
UV_ART| Characteristic | N | HR | 95% CI |
|---|---|---|---|
| Age | 1,957 | ||
|     25-35 | — | — | |
| Â Â Â Â 18-25 | 0.76 | 0.43, 1.35 | |
| Â Â Â Â 35-45 | 1.01 | 0.77, 1.31 | |
| Â Â Â Â >45 | 1.36 | 1.04, 1.78 | |
| Sex (Male) | 1,957 | 1.11 | 0.90, 1.37 |
| CD4 | 1,957 | ||
|     101 - 199 | — | — | |
| Â Â Â Â 51 - 100 | 1.71 | 1.31, 2.22 | |
| Â Â Â Â 0 - 50 | 2.59 | 2.01, 3.33 | |
| Previous TB | 1,957 | 1.56 | 1.24, 1.97 |
| Incident TB | 1,957 | 1.17 | 0.92, 1.50 |
| Current TB | 1,957 | 2.18 | 1.71, 2.79 |
| Previous Cryptococcosis | 1,957 | 0.96 | 0.49, 1.85 |
| Viral Load at Enumeration | 1,957 | ||
|     < 100 | — | — | |
| Â Â Â Â 100 -1000 | 1.00 | 0.62, 1.61 | |
| Â Â Â Â > 1000 | 1.50 | 1.17, 1.91 | |
| Â Â Â Â None | 1.18 | 0.85, 1.65 | |
| Enumerated in Hospital | 1,957 | 3.57 | 2.88, 4.42 |
| Abbreviations: CI = Confidence Interval, HR = Hazard Ratio | |||
UV_Dis| Characteristic | N | HR | 95% CI |
|---|---|---|---|
| Age | 4,467 | ||
|     25-35 | — | — | |
| Â Â Â Â 18-25 | 1.36 | 0.99, 1.86 | |
| Â Â Â Â 35-45 | 1.23 | 1.02, 1.49 | |
| Â Â Â Â >45 | 1.82 | 1.47, 2.25 | |
| Sex (Male) | 4,467 | 1.24 | 1.06, 1.46 |
| CD4 | 4,467 | ||
|     101 - 199 | — | — | |
| Â Â Â Â 51 - 100 | 1.74 | 1.40, 2.17 | |
| Â Â Â Â 0 - 50 | 3.58 | 2.97, 4.31 | |
| Previous TB | 4,467 | 1.75 | 1.49, 2.06 |
| Incident TB | 4,467 | 1.29 | 1.08, 1.54 |
| Current TB | 4,467 | 1.91 | 1.62, 2.25 |
| Previous Cryptococcosis | 4,467 | 1.36 | 0.68, 2.73 |
| Viral Load at Enumeration | 4,467 | ||
|     < 100 | — | — | |
| Â Â Â Â 100 -1000 | 1.08 | 0.63, 1.86 | |
| Â Â Â Â > 1000 | 1.68 | 1.16, 2.41 | |
| Â Â Â Â None | 1.10 | 0.78, 1.55 | |
| Enumerated in Hospital | 4,467 | 2.73 | 2.29, 3.25 |
| Abbreviations: CI = Confidence Interval, HR = Hazard Ratio | |||
UV_Naive| Characteristic | N | HR | 95% CI |
|---|---|---|---|
| Age | 4,728 | ||
|     25-35 | — | — | |
| Â Â Â Â 18-25 | 0.97 | 0.65, 1.45 | |
| Â Â Â Â 35-45 | 1.69 | 1.33, 2.16 | |
| Â Â Â Â >45 | 2.41 | 1.85, 3.13 | |
| Sex (Male) | 4,728 | 1.46 | 1.19, 1.78 |
| CD4 | 4,728 | ||
|     101 - 199 | — | — | |
| Â Â Â Â 51 - 100 | 1.49 | 1.16, 1.92 | |
| Â Â Â Â 0 - 50 | 2.65 | 2.11, 3.33 | |
| Previous TB | 4,728 | 1.73 | 1.28, 2.32 |
| Incident TB | 4,728 | 1.27 | 0.98, 1.64 |
| Current TB | 4,728 | 2.30 | 1.89, 2.80 |
| Enumerated in Hospital | 4,728 | 2.62 | 2.06, 3.34 |
| Abbreviations: CI = Confidence Interval, HR = Hazard Ratio | |||
Combining all models into one single table
MV_1 <- tbl_regression(surv_reg, exponentiate = TRUE,pvalue_fun = ~style_pvalue(.x, digits = 2),) %>%
bold_labels()
MV_1 <- MV_1 %>%
modify_table_body(
~ .x %>% select(-p.value)
)
MV_2 <- tbl_regression(surv_ART1, exponentiate = TRUE,pvalue_fun = ~style_pvalue(.x, digits = 2),) %>%
bold_labels()
MV_2 <- MV_2 %>%
modify_table_body(
~ .x %>% select(-p.value)
)
MV_3 <- tbl_regression(surv_ART2, exponentiate = TRUE,pvalue_fun = ~style_pvalue(.x, digits = 2),) %>%
bold_labels()
MV_3 <- MV_3 %>%
modify_table_body(
~ .x %>% select(-p.value)
)
MV_4 <- tbl_regression(surv_Dis, exponentiate = TRUE,pvalue_fun = ~style_pvalue(.x, digits = 2),) %>%
bold_labels()
MV_4 <- MV_4 %>%
modify_table_body(
~ .x %>% select(-p.value)
)
MV_5 <- tbl_regression(surv_Naive, exponentiate = TRUE,pvalue_fun = ~style_pvalue(.x, digits = 2),) %>%
bold_labels()
MV_5 <- MV_5 %>%
modify_table_body(
~ .x %>% select(-p.value)
)
theme_gtsummary_compact()
tbl_merge(
tbls = list(UV_all, MV_1, UV_ART, MV_2, MV_3, UV_Dis, MV_4, UV_Naive, MV_5), # combine
tab_spanner = c("**Univariable**",
"**Overall**",
"**Univariable**",
"**On ART 1**",
"**On ART 2**",
"**Univariable**",
"**Disengaged**",
"**Univariable**",
"**ART Naive**"))| Characteristic |
Univariable
|
Overall
|
Univariable
|
On ART 1
|
On ART 2
|
Univariable
|
Disengaged
|
Univariable
|
ART Naive
|
|||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N | HR | 95% CI | HR | 95% CI | N | HR | 95% CI | HR | 95% CI | HR | 95% CI | N | HR | 95% CI | HR | 95% CI | N | HR | 95% CI | HR | 95% CI | |
| Age | 11,152 | 1,957 | 4,467 | 4,728 | ||||||||||||||||||
|     25-35 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | ||||
| Â Â Â Â 18-25 | 1.03 | 0.82, 1.29 | 1.19 | 0.95, 1.50 | 0.76 | 0.43, 1.35 | 0.75 | 0.42, 1.34 | 0.78 | 0.44, 1.38 | 1.36 | 0.99, 1.86 | 1.54 | 1.12, 2.12 | 0.97 | 0.65, 1.45 | 1.16 | 0.77, 1.74 | ||||
| Â Â Â Â 35-45 | 1.36 | 1.19, 1.55 | 1.20 | 1.05, 1.37 | 1.01 | 0.77, 1.31 | 1.00 | 0.77, 1.31 | 1.00 | 0.76, 1.30 | 1.23 | 1.02, 1.49 | 1.10 | 0.90, 1.34 | 1.69 | 1.33, 2.16 | 1.57 | 1.23, 2.01 | ||||
| Â Â Â Â >45 | 1.94 | 1.68, 2.23 | 1.75 | 1.52, 2.03 | 1.36 | 1.04, 1.78 | 1.48 | 1.12, 1.96 | 1.41 | 1.07, 1.86 | 1.82 | 1.47, 2.25 | 1.68 | 1.34, 2.11 | 2.41 | 1.85, 3.13 | 2.25 | 1.73, 2.94 | ||||
| ART | 11,152 | |||||||||||||||||||||
|     On ART | — | — | — | — | ||||||||||||||||||
| Â Â Â Â LTFU | 0.84 | 0.74, 0.96 | 0.82 | 0.72, 0.94 | ||||||||||||||||||
| Â Â Â Â ART Naive | 0.49 | 0.43, 0.57 | 0.60 | 0.51, 0.70 | ||||||||||||||||||
| Sex (Male) | 11,152 | 1.22 | 1.09, 1.36 | 1.04 | 0.93, 1.16 | 1,957 | 1.11 | 0.90, 1.37 | 0.95 | 0.76, 1.19 | 1.02 | 0.82, 1.27 | 4,467 | 1.24 | 1.06, 1.46 | 0.97 | 0.82, 1.16 | 4,728 | 1.46 | 1.19, 1.78 | 1.26 | 1.03, 1.54 |
| CD4 | 11,152 | 1,957 | 4,467 | 4,728 | ||||||||||||||||||
|     101 - 199 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | ||||||
| Â Â Â Â 51 - 100 | 1.60 | 1.39, 1.84 | 1.62 | 1.40, 1.86 | 1.71 | 1.31, 2.22 | 1.70 | 1.30, 2.21 | 1.74 | 1.40, 2.17 | 1.71 | 1.38, 2.13 | 1.49 | 1.16, 1.92 | 1.46 | 1.13, 1.88 | ||||||
| Â Â Â Â 0 - 50 | 2.94 | 2.60, 3.33 | 2.90 | 2.56, 3.29 | 2.59 | 2.01, 3.33 | 2.59 | 2.01, 3.35 | 3.58 | 2.97, 4.31 | 3.41 | 2.82, 4.11 | 2.65 | 2.11, 3.33 | 2.45 | 1.95, 3.09 | ||||||
| Previous TB | 11,152 | 2.00 | 1.80, 2.23 | 1.47 | 1.30, 1.67 | 1,957 | 1.56 | 1.24, 1.97 | 1.45 | 1.14, 1.83 | 1.50 | 1.19, 1.90 | 4,467 | 1.75 | 1.49, 2.06 | 1.50 | 1.27, 1.77 | 4,728 | 1.73 | 1.28, 2.32 | 1.54 | 1.14, 2.07 |
| Incident TB | 11,152 | 1.34 | 1.19, 1.52 | 1,957 | 1.17 | 0.92, 1.50 | 4,467 | 1.29 | 1.08, 1.54 | 4,728 | 1.27 | 0.98, 1.64 | ||||||||||
| Current TB | 11,152 | 1.88 | 1.68, 2.10 | 1,957 | 2.18 | 1.71, 2.79 | 4,467 | 1.91 | 1.62, 2.25 | 4,728 | 2.30 | 1.89, 2.80 | ||||||||||
| Previous Cryptococcosis | 11,152 | 1.40 | 0.87, 2.25 | 0.88 | 0.55, 1.43 | 1,957 | 0.96 | 0.49, 1.85 | 0.82 | 0.42, 1.59 | 0.87 | 0.45, 1.68 | 4,467 | 1.36 | 0.68, 2.73 | 1.02 | 0.51, 2.06 | |||||
| Viral Load at Enumeration | 11,152 | 1,957 | 4,467 | |||||||||||||||||||
|     < 100 | — | — | — | — | — | — | — | — | ||||||||||||||
| Â Â Â Â 100 -1000 | 0.97 | 0.70, 1.35 | 1.00 | 0.62, 1.61 | 1.04 | 0.64, 1.68 | 1.08 | 0.63, 1.86 | ||||||||||||||
| Â Â Â Â > 1000 | 1.53 | 1.27, 1.85 | 1.50 | 1.17, 1.91 | 1.51 | 1.18, 1.92 | 1.68 | 1.16, 2.41 | ||||||||||||||
| Â Â Â Â None | 0.78 | 0.65, 0.93 | 1.18 | 0.85, 1.65 | 1.25 | 0.89, 1.76 | 1.10 | 0.78, 1.55 | ||||||||||||||
| Enumerated in Hospital | 11,152 | 3.11 | 2.77, 3.49 | 1,957 | 3.57 | 2.88, 4.42 | 4,467 | 2.73 | 2.29, 3.25 | 4,728 | 2.62 | 2.06, 3.34 | ||||||||||
| Abbreviations: CI = Confidence Interval, HR = Hazard Ratio | ||||||||||||||||||||||
Diagnostics and alterantive considerations for overall CPH model
Diagnostics
## chisq df p
## ART 15.393 2 0.00045
## `Sex (Male)` 0.903 1 0.34186
## Age 23.245 3 3.6e-05
## CD4 14.387 2 0.00075
## `Previous TB` 1.627 1 0.20212
## `Previous Cryptococcosis` 0.288 1 0.59180
## GLOBAL 58.727 10 6.3e-09## `geom_smooth()` using formula = 'y ~ x'
Age
KM Curves and log-log curves
## Warning: Removed 285 rows containing missing values or values outside the scale range
## (`geom_step()`).## Warning: Removed 269 rows containing missing values or values outside the scale range
## (`geom_point()`).## Warning: Removed 1 row containing missing values or values outside the scale range
## (`geom_text()`).## Warning: Removed 285 rows containing missing values or values outside the scale range
## (`geom_step()`).## Warning: Removed 269 rows containing missing values or values outside the scale range
## (`geom_point()`).## Warning: Removed 1 row containing missing values or values outside the scale range
## (`geom_text()`).
Schoenfield residuals
## Warning in regularize.values(x, y, ties, missing(ties), na.rm = na.rm):
## collapsing to unique 'x' values
Summary of Age:
The age groups of 18-25 and 25-35 are clearly statstically significant as they cross. However, they seem so co-linear that it look like their hazard curves are almost the same in which case the crossing may not be a major issue.
CD4 count
KM Curves and log-log curves
## Warning: Removed 550 rows containing missing values or values outside the scale range
## (`geom_step()`).## Warning: Removed 519 rows containing missing values or values outside the scale range
## (`geom_point()`).## Warning: Removed 1 row containing missing values or values outside the scale range
## (`geom_text()`).## Warning: Removed 550 rows containing missing values or values outside the scale range
## (`geom_step()`).## Warning: Removed 519 rows containing missing values or values outside the scale range
## (`geom_point()`).## Warning: Removed 1 row containing missing values or values outside the scale range
## (`geom_text()`).
Schoenfield residuals
## Warning in regularize.values(x, y, ties, missing(ties), na.rm = na.rm):
## collapsing to unique 'x' values
Summary of CD4
CD4 doesn’t appear to violate the PH assumption signficiantly. Likely statistically significant because of the very beginning and end.
ART
KM Curves and log-log curves
## Warning: Removed 288 rows containing missing values or values outside the scale range
## (`geom_step()`).## Warning: Removed 274 rows containing missing values or values outside the scale range
## (`geom_point()`).## Warning: Removed 1 row containing missing values or values outside the scale range
## (`geom_text()`).## Warning: Removed 288 rows containing missing values or values outside the scale range
## (`geom_step()`).## Warning: Removed 274 rows containing missing values or values outside the scale range
## (`geom_point()`).## Warning: Removed 1 row containing missing values or values outside the scale range
## (`geom_text()`).
Schoenfield residuals
## Warning in regularize.values(x, y, ties, missing(ties), na.rm = na.rm):
## collapsing to unique 'x' values
Summary of ART
ART violates the PH assumption mostly in the early time period - this is best seen in the KM curves where the on ART group has an acute mortality risk that is much steeper than the other two groups. Various modelling strategies to overcome the violation of the PH assumption are subsequently evaluated. However, after evaluating all strategies and review with all authors it was concluded that ultimately the purpose of modelling in this study is not to get accurate estimates of mortality risk as the three ART groups reflect three different populations each confounded by important selection biases thus attempting to estimate accurate hazard/risk ratios over time would be attempting to approximate a false level of precision. Thus a decision was made to stick with CPH modelling while clearly understanding the limitations of it in this context.
Managing violations of the PH assumption
four approaches to managing the violation of the CPH assumption within ART were considered. 1- stratifying by ART category (implemented in final modelling) 2- estimating a fully parametric model 3- emplyoing step functions 4- using time transformations The most successful approaches were stratification and the use of an stpm2 model with 3 knots, 4 splines and a PH link function.
However as discussed above it was ultimately decided to go with a CPH model as 1) the violation of the PH assumption is limited to the first time period for reasons discussed in the discussion and 2) the purpose of modelling in this study is hypothesis generating rather than attempting to accurately estimate risk magnitudes.
Fitting stpm2 model
library(rstpm2)## Loading required package: splines##
## Attaching package: 'rstpm2'## The following object is masked from 'package:survival':
##
## colonfit <- stpm2(Surv(Time, Censored) ~ ART, data = Model_Data, df=4)
eform(fit)[1:3,]## exp(beta) 2.5 % 97.5 %
## (Intercept) 0.0008561144 0.0008112656 0.0009024998
## ARTLTFU 0.8346362570 0.7700323597 0.9027361177
## ARTART Naive 0.4877151354 0.4411838490 0.5373987454# Fit Cox proportional hazards model
cox.fit <- coxph(Surv(Time, Censored) ~ ART, data = Model_Data)
# Create survival curves (baseline survival function)
surv_curve <- survfit(cox.fit)
# Extract baseline hazard function
baseline_hazard <- -diff(surv_curve$surv) / diff(surv_curve$time)
# Plot the baseline hazard function
plot(surv_curve$time[-1], baseline_hazard, type = 'l',
xlab = 'Time', ylab = 'Baseline Hazard',
main = 'Baseline Hazard Function of Cox Model',
ylim = c(0,0.005))
plot(fit, newdata=data.frame(ART = "On ART"),
xlab="Time since enumeration (days)", main = "Univaraite flexible PH parametric model of ART plotted against KM")
lines(fit, newdata=data.frame(ART = "LTFU"), lty=2)
lines(fit, newdata=data.frame(ART = "ART Naive"), lty=3)
lines(survfit(Surv(Time, Censored)~ART, data=Model_Data), col="red", lty=1:3)
legend("topright", c("PH On ART","PH Disengaged","PH ART Naive",
"KM On ART","KM Disengaged","KM ART Naive"),
lty=1:3, col=c("black","black","black","red","red","red"))
Plotting hazard function
predART <- predict(fit, newdata=data.frame(ART = c("On ART", "LTFU", "ART Naive")),
type = "hazard", grid=TRUE, full=TRUE, se.fit=TRUE)
#Transform
predART <- transform(predART,ART=factor(ART,labels=c("On ART","Disengaged", "Naive")))
#Plot
ggplot(predART, aes(x=Time, y=Estimate, ymin=lower, ymax=upper, fill = ART)) +
xlab("Time since enumeration") +
ylab("Hazard") +
geom_ribbon() +
geom_line()
Additional stats and calculations
# Evaluating re-initiating ART
Disengaged <- subset(Cohort, ART_Exp_C == "On ART")
table(Disengaged$Censored)##
## 0 1
## 2033 346summary(Disengaged$Enumeration_CD4)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0 74.0 126.0 117.2 166.0 199.0Disengaged_Time <- subset(Disengaged, Reinitiated == 1)
Disengaged_Time_0 <- subset(Disengaged_Time, Time_Reinitiation == 0)
Disengaged_14 <- subset(Disengaged, Time_Reinitiation < 15)
# Re-initiated in Hospital
table(Disengaged$Hospital_Enumerations)##
## 0 1
## 1813 566Hospital_Disengaged <- subset(Disengaged, Hospital_Enumerations == 1)
Hospital_Disengaged <- subset(Hospital_Disengaged, Linked == 1)
Disengaged_H_60 <- subset(Hospital_Disengaged, Time_Reinitiation < 61)
Disengaged_H_D_60 <- subset(Hospital_Disengaged, Time_Reinitiation > 60 | Reinitiated == 0)
Disengaged_H_D_60_Less <- subset(Disengaged_H_D_60, Time < 61)
Disengaged_C <- subset(Disengaged_H_D_60, Time > 60)
table(Disengaged_C$Prevalent_Crypto)## < table of extent 0 >table(Disengaged_C$Current_TB_PHDC)## < table of extent 0 >Disengaged_C <- subset(Disengaged_C, Prevalent_Crypto == 0 & Current_TB_PHDC == 0)
table(Disengaged_C$Reinitiated)## < table of extent 0 >#Re-initiated in PHC
table(Disengaged$Hospital_Enumerations)##
## 0 1
## 1813 566PHC_Disengaged <- subset(Disengaged, Hospital_Enumerations == 0)
PHC_Disengaged <- subset(PHC_Disengaged, Linked == 1)
Disengaged_P_60 <- subset(PHC_Disengaged, Time_Reinitiation < 61)
Disengaged_P_D_60 <- subset(PHC_Disengaged, Time_Reinitiation > 60 | Reinitiated == 0)
Disengaged_P_D_60_Less <- subset(Disengaged_P_D_60, Time < 61)
Disengaged_P_C <- subset(Disengaged_P_D_60, Time > 60)
table(Disengaged_P_C$Prevalent_Crypto)## < table of extent 0 >table(Disengaged_P_C$Current_TB_PHDC)## < table of extent 0 >Disengaged_P_C <- subset(Disengaged_P_C, Prevalent_Crypto == 0 & Current_TB_PHDC == 0)
table(Disengaged_P_C$Reinitiated)## < table of extent 0 ># Evaulating First Starting ART
Naive <- subset(Cohort, ART_Exp_C == "Naive")
table(Naive$Censored)##
## 0 1
## 5290 395summary(Naive$Enumeration_CD4)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0 56.0 108.0 105.9 157.0 199.0Naive_Time <- subset(Naive, Subs_Initiated == 1)
Naive_Time_0 <- subset(Naive_Time, Time_To_First_ART == 0)
Naive_14 <- subset(Naive, Time_To_First_ART < 15)
#First ART in Hospital
table(Naive$Hospital_Enumerations)##
## 0 1
## 5043 642Hospital_Naive <- subset(Naive, Hospital_Enumerations == 1)
Hospital_Naive <- subset(Hospital_Naive, Linked == 1)
Naive_H_60 <- subset(Hospital_Naive, Time_To_First_ART < 61)
Naive_H_D_60 <- subset(Hospital_Naive, Time_To_First_ART > 60 | Subs_Initiated == 0)
Naive_H_D_60_Less <- subset(Naive_H_D_60, Time < 61)
Naive_C <- subset(Naive_H_D_60, Time > 60)
table(Naive_C$Prevalent_Crypto)##
## 0 1
## 112 6table(Naive_C$Current_TB_PHDC)##
## 0 1
## 66 52Naive_C <- subset(Naive_C, Prevalent_Crypto == 0 & Current_TB_PHDC == 0)
table(Naive_C$Subs_Initiated)##
## 0 1
## 18 44#First ART in PHC
table(Naive$Hospital_Enumerations)##
## 0 1
## 5043 642PHC_Naive <- subset(Naive, Hospital_Enumerations == 0)
PHC_Naive <- subset(PHC_Naive, Linked == 1)
Naive_P_60 <- subset(PHC_Naive, Time_To_First_ART < 61)
Naive_P_D_60 <- subset(PHC_Naive, Time_To_First_ART > 60 | Subs_Initiated == 0)
Naive_P_D_60_Less <- subset(Naive_P_D_60, Time < 61)
Naive_P_C <- subset(Naive_P_D_60, Time > 60)
table(Naive_P_C$Prevalent_Crypto)##
## 0 1
## 634 11table(Naive_P_C$Current_TB_PHDC)##
## 0 1
## 477 168Naive_P_C <- subset(Naive_P_C, Prevalent_Crypto == 0 & Current_TB_PHDC == 0)
table(Naive_P_C$Subs_Initiated)##
## 0 1
## 74 397table(Disengaged_C$Reinitiated, Disengaged_C$Censored)## < table of extent 0 x 0 >table(Disengaged_P_C$Reinitiated, Disengaged_P_C$Censored)## < table of extent 0 x 0 >table(Naive_C$Subs_Initiated, Naive_C$Censored)##
## 0 1
## 0 16 2
## 1 40 4table(Naive_P_C$Subs_Initiated, Naive_P_C$Censored)##
## 0 1
## 0 68 6
## 1 370 27#One year mortality
Cohort_Linked <- subset(Cohort, Linked == 1)
Cohort_Year <- subset(Cohort_Linked, Enumeration_date <= "2020-03-30")
Cohort_Died_More <- subset(Cohort_Year, Censored==1 & Time >365)
Cohort_Died_More$More_than_Year <- 1
Cohort_Year$More_than_Year <-
Cohort_Died_More$More_than_Year[match(Cohort_Year$study_id, Cohort_Died_More$study_id)]
Cohort_Year$More_than_Year[is.na(Cohort_Year$More_than_Year)] = 0
Cohort_Year <- subset(Cohort_Year, More_than_Year < 1)
table(Cohort_Year$Censored)##
## 0 1
## 8323 646Cohort_Censored <- subset(Linked_Data, Mortality == "Died")
plot(Cohort_Censored$Time)
Cen_Plot <- ggdensity(Cohort_Censored, x = "Time",
add = "median",
color = "ART", fill = "ART",
palette = "npg",
title = "Density plot of time to death by enumeration ART status")
Cen_Plot
Disengaged_60 <- subset(Disengaged, Time_Reinitiation < 61)
Naive_60 <- subset(Naive, Time_To_First_ART < 61)